{"id":24127,"date":"2026-09-24T13:19:09","date_gmt":"2026-09-24T11:19:09","guid":{"rendered":"https:\/\/www.neuraldesigner.com\/?post_type=learning&#038;p=24127"},"modified":"2026-10-06T19:13:22","modified_gmt":"2026-10-06T17:13:22","slug":"melanoma-image-classification","status":"publish","type":"learning","link":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/","title":{"rendered":"Melanoma classification from dermoscopic images"},"content":{"rendered":"<style>\n.nds{box-sizing:border-box;--nds-code-bg:#f3f7fa;--nds-code-fg:#12354b;width:100%;margin-left:0;margin-right:0;padding:22px 0 14px;background:#fff;color:#1b2635;font-family:\"Outfit\",\"Roboto\",Arial,sans-serif}\n.nds *{box-sizing:border-box}.nds-wrap{width:min(calc(100% - 48px),1160px);margin:0 auto}.nds a{text-decoration:none;color:#2d799f}\n.nds-executive{margin:0 0 34px;padding:32px;border-radius:20px;}.nds-executive h2{margin:0 0 12px;font-size:30px}.nds-executive p{font-size:18px;line-height:1.55}\n.nds-kpis{display:grid;grid-template-columns:repeat(4,minmax(0,1fr));gap:14px;margin-top:22px}.nds-kpi{padding:18px;border:1px solid rgba(255,255,255,.18);border-radius:14px}.nds-kpi strong{display:block;font-size:25px}.nds-kpi span{font-size:13px}\n.nds-actions,.nds-audience,.nds-downloads{display:flex;flex-wrap:wrap;justify-content:center;gap:12px;margin:22px 0}\n.nds-toc{display:flex;flex-wrap:wrap;justify-content:center;gap:10px;margin:0 0 42px;padding:0;list-style:none}.nds-toc a,.nds-audience span{display:block;padding:9px 14px;border-radius:20px;background:#e9f2f8;color:#12354b;font-weight:600}\n.nds-card{margin:0 0 54px;scroll-margin-top:90px}.nds-card h2{margin:0 0 20px;padding-bottom:12px;border-bottom:1px solid #dbe5ec;color:#001233;font-size:24px}.nds-card p,.nds-card li{font-size:16.5px;line-height:1.6}.nds-card img{display:block;width:auto;max-width:min(640px,100%);height:auto;margin:24px auto;border-radius:12px}.nds-card img.nds-architecture{width:min(1000px,100%);max-width:100%}\n.nds-value-grid{display:grid;grid-template-columns:repeat(3,minmax(0,1fr));gap:16px;margin:22px 0}.nds-value{padding:20px;border:1px solid #dce8ef;border-radius:14px;background:#f8fbfd}.nds-note{margin:20px 0;padding:18px 20px;border-left:4px solid #56a1c8;border-radius:0 12px 12px 0;background:#f6fafc}.nds-note--warning,.nds-use-boundary{border-left-color:#d79a29;background:#fff9ed}\n.nds-table-scroll{width:100%;max-width:100%;overflow-x:auto}.nds-card table{width:auto;max-width:100%;margin:24px auto;border-collapse:collapse;background:#fff}.nds-card th,.nds-card td{padding:11px 16px;border-bottom:1px solid #e2e9ee}.nds-card thead th{background:#12354b!important;color:#fff!important}.nds-card tbody th{background:#eaf2f6!important;color:#12354b!important;text-align:left}.nds-card tbody tr:nth-child(even) th{background:#f4f8fa!important}.nds-card tbody td{color:#33424f!important}\nbody.nd-design-safe .nd-native-article__content .nds-table-scroll table{display:table!important;width:auto!important;min-width:0!important;max-width:100%!important;margin:24px auto!important;overflow:visible!important}\n.nds-figure-grid{display:grid;grid-template-columns:repeat(auto-fit,minmax(min(100%,520px),600px));justify-content:center;gap:24px;margin:24px 0}.nds-figure-grid figure,.nds-centered-figure,.nds-architecture-figure,.nds-interactive-figure{width:100%;max-width:100%;margin:0 auto!important;padding:0;border:0;background:transparent}.nds-centered-figure{width:min(720px,100%)}.nds-architecture-figure{width:min(1120px,100%)}.nds-interactive-figure{width:min(920px,100%)}.nds-roc-figure{width:min(620px,100%)}.nds-architecture-link{display:block}.nds-figure-grid img,.nds-card figure.nds-centered-figure>img,.nds-card figure.nds-architecture-figure img{display:block!important;width:100%!important;max-width:100%!important;height:auto!important;margin:0 auto 12px!important}.nds-interactive-chart{width:100%;height:clamp(390px,48vw,560px);overflow:hidden;border:1px solid #dce5ea;border-radius:10px;background:#fff}.nds-interactive-figure--dataset .nds-interactive-chart{height:clamp(680px,72vw,780px)}.nds-card figcaption{color:#405361;font-size:14px;line-height:1.45}.nds-interactive-figure figcaption{margin-top:10px;text-align:left}\n.nds-comparison-grid{grid-column:1\/-1;display:grid;grid-template-columns:minmax(0,1fr);justify-self:center;gap:28px;width:100%;margin:28px 0}.nds-comparison-grid>.nds-comparison-item{min-width:0}.nds-comparison-grid>br,.nds-comparison-grid>p:empty{display:none!important}.nds-comparison-grid .nds-interactive-figure{display:flex;flex-direction:column;width:100%}.nds-comparison-grid .nds-interactive-chart{height:clamp(420px,42vw,520px)}@media(min-width:1180px){.nds-comparison-grid{grid-template-columns:repeat(2,minmax(0,1fr));align-items:start;width:100%;max-width:100%}.nds-comparison-grid .nds-interactive-chart{height:500px}}\n.nds pre,.nds .wp-block-code,.nds pre[class*=\"language-\"]{display:block;max-width:100%;overflow:auto;padding:18px 20px!important;border:1px solid #d5e2e9!important;border-radius:12px!important;background:var(--nds-code-bg)!important;color:var(--nds-code-fg)!important;box-shadow:none!important;text-shadow:none!important;white-space:pre}\n.nds pre code,.nds pre code *,.nds .wp-block-code code,.nds .wp-block-code code *{border:0!important;background:transparent!important;color:var(--nds-code-fg)!important;box-shadow:none!important;text-shadow:none!important;font-family:Consolas,\"Liberation Mono\",monospace!important;font-size:14px;line-height:1.55}\n.nds :not(pre)>code{padding:.08em .32em;border-radius:4px;background:#e6eef3!important;color:#12354b!important}.nds .nds-expression{max-width:100%;overflow:auto;margin:22px auto;padding:18px 20px;border:1px solid #d5e2e9;border-radius:12px;background:#fff!important;color:#12354b!important;text-align:center}.nds .nds-expression math,.nds .nds-expression mjx-container,.nds .nds-expression .MathJax{background:transparent!important;color:#12354b!important}\n.nds .nds-calculator-grid>br,.nds .nds-calculator-grid>p:empty{display:none!important}\n@media(max-width:820px){.nds-kpis{grid-template-columns:repeat(2,minmax(0,1fr))}.nds-value-grid{grid-template-columns:1fr}}@media(max-width:620px){.nds{padding:12px 0}.nds-kpis,.nds-figure-grid{grid-template-columns:1fr}.nds-executive{padding:24px 20px}.nds-card table{font-size:13px}.nds-card th,.nds-card td{padding:9px 8px}}\n\/* nd-shared-components:start *\/html body.nds-shared:not(.wp-admin), html body.nds-shared:not(.wp-admin) :is(.site-main,.site-content,.entry-content,.page-content,.nd-native-article,.nd-native-article__content,.nd-home-page,.nd-product-page,.nd-company-page,.nd-contact-page,.nd-downloads-page,.nd-account-page,.nd-plain-page,.nd-blog-page,.ndlh-page,.ndhub,.ndth,.ndts,.ndareas,.ndarea,.ndeu), html body:not(.wp-admin) :is(.ndb,.nds,.ndg,.nd-page-surface){background-color:#fff!important;background-image:none!important} :is(.ndb,.nds) :is(.ndb-executive,.nds-executive,.ndb-highlight){background:#f4f7f9;color:#193645} :is(.ndb,.nds) :is(.ndb-executive,.nds-executive,.ndb-highlight) :is(h2,h3){color:#193645} :is(.ndb,.nds) :is(.ndb-executive,.nds-executive,.ndb-highlight) p{color:#52636f} :is(.ndb,.nds) :is(.ndb-kpi,.nds-kpi){background:#245e80;color:#fff} :is(.ndb,.nds) :is(.ndb-kpi,.nds-kpi) strong{color:#fff} :is(.ndb,.nds) :is(.ndb-kpi,.nds-kpi) span{color:#d9edf7} :is(.ndb,.nds) :is(.aui-button,.ndb-actions a,.ndb-downloads a,.nds-actions a,.nds-downloads a){box-sizing:border-box;display:inline-flex;align-items:center;justify-content:center;gap:8px;min-width:0;min-height:48px;max-width:100%;height:auto;padding:11px 20px;background:#176b91;color:#fff;border:1px solid #176b91;border-radius:6px;font-family:Roboto,\"Segoe UI\",Arial,sans-serif;font-size:16px;font-weight:700;line-height:1.5;letter-spacing:normal;text-align:center;text-decoration:none;text-transform:none;white-space:normal;overflow-wrap:anywhere;box-shadow:none;cursor:pointer} :is(.ndb,.nds) :is(.aui-button,.ndb-actions a,.ndb-downloads a,.nds-actions a,.nds-downloads a):is(:hover,:focus-visible){background:#125574;color:#fff;border-color:#125574} :is(.ndb,.nds) :is(.aui-button,.ndb-actions a,.ndb-downloads a,.nds-actions a,.nds-downloads a):focus-visible{outline:3px solid #176b91;outline-offset:4px} :is(.ndb,.nds) .aui-button.aui-button--secondary{background:#fff;color:#176b91;border-color:#176b91} :is(.ndb,.nds) .aui-button.aui-button--secondary:is(:hover,:focus-visible){background:#f4f7f9;color:#125574;border-color:#125574} :is(.ndb,.nds) .aui-button:is(:disabled,[aria-disabled=\"true\"],.disabled){opacity:.55;cursor:not-allowed} @media(forced-colors:active){:is(.ndb,.nds) .aui-button{border-color:ButtonText}:is(.ndb,.nds) .aui-button:focus-visible{outline-color:Highlight}} html body :is(.nd-native-article__content,.entry-content,.page-content,.ndb,.nds,.ndg) :is(.nd-table-scroll,.ndb-table-scroll,.nds-table-scroll,.ndg-table-wrap,.ndb-table-wrap,.ndb-confusion-wrap,.wp-block-table){box-sizing:border-box;display:block!important;width:100%!important;min-width:0!important;max-width:100%!important;padding:0!important;background:transparent!important;border:0!important;box-shadow:none!important;overflow-x:auto!important;overscroll-behavior-inline:contain;-webkit-overflow-scrolling:touch} html body :is(.nd-native-article__content,.entry-content,.page-content,.ndb,.nds,.ndg) :is(.nd-table-scroll,.ndb-table-scroll,.nds-table-scroll,.ndg-table-wrap,.ndb-table-wrap,.ndb-confusion-wrap,.wp-block-table)>table, html body :is(.nd-native-article__content,.entry-content,.page-content,.ndb,.nds,.ndg) table.nd-content-table{display:table!important;width:auto!important;min-width:0!important;max-width:100%!important;margin:24px auto!important;float:none!important;table-layout:auto!important;overflow:visible!important} html body :is(.nd-native-article__content,.entry-content,.page-content,.ndb,.nds,.ndg) .nd-table-scroll:focus-visible{outline:3px solid #176b91;outline-offset:3px} html body.nds-shared:not(.wp-admin) .aui-button.aui-button, html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button{box-sizing:border-box;min-height:48px!important;height:auto!important;padding:11px 20px!important;border:1px solid #176b91!important;border-radius:6px!important;font-family:Roboto,\"Segoe UI\",Arial,sans-serif!important;font-size:16px!important;font-weight:700!important;line-height:1.5!important;color:#fff!important;-webkit-text-fill-color:#fff!important;background:#176b91!important;text-shadow:none!important;box-shadow:none!important;transform:none!important;translate:none!important;scale:none!important;rotate:none!important;transition:none!important;animation:none!important} html body.nds-shared:not(.wp-admin) .aui-button.aui-button:is(:hover,:focus-visible), html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button:is(:hover,:focus-visible){color:#fff!important;-webkit-text-fill-color:#fff!important;background:#125574!important;border-color:#125574!important} html body.nds-shared:not(.wp-admin) .aui-button.aui-button:is(:hover,:focus,:focus-visible), html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button:is(:hover,:focus,:focus-visible){box-shadow:none!important;transform:none!important;translate:none!important;scale:none!important;rotate:none!important;transition:none!important;animation:none!important} html body.nds-shared:not(.wp-admin) .aui-button.aui-button.aui-button--secondary, html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button.aui-button--secondary{color:#176b91!important;-webkit-text-fill-color:#176b91!important;background:#fff!important;border-color:#176b91!important} html body.nds-shared:not(.wp-admin) .aui-button.aui-button.aui-button--secondary:is(:hover,:focus-visible), html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button.aui-button--secondary:is(:hover,:focus-visible){color:#125574!important;-webkit-text-fill-color:#125574!important;background:#f4f7f9!important;border-color:#125574!important} html body.nds-shared:not(.wp-admin) .aui-button.aui-button:focus-visible, html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button:focus-visible{outline:3px solid #176b91!important;outline-offset:4px!important} html body.nds-shared:not(.wp-admin) .aui-button.aui-button :is(span,strong,b,em,i), html body :is(.ndb,.nds,.nd-page-surface) .aui-button.aui-button :is(span,strong,b,em,i){color:inherit!important;-webkit-text-fill-color:inherit!important} @media(forced-colors:active){html body .aui-button.aui-button{border-color:ButtonText!important}html body .aui-button.aui-button:focus-visible{outline-color:Highlight!important}} html body{--nd-content-width:1160px;--nd-content-gutter:24px} @media(max-width:640px){html body{--nd-content-gutter:20px}} html body.nds-shared .aui{--aui-width:var(--nd-content-width);--aui-gutter:var(--nd-content-gutter)} html body.nds-shared :is(.aui-container,.nd-native-article__content,.nds-article-content,.nd-home-wrap,.nd-product-wrap,.nd-company-main,.ndlh-wrap,.nd-blog-main,.ndhub-wrap,.ndth-wrap,.ndts-wrap,.ndareas-wrap,.ndarea-wrap,.ndeu-wrap,.nd-contact-wrap,.nd-downloads-wrap,.nd-plain-wrap,.nd-account-hero-inner,.nd-account-main), html body :is(.nd-content-container,.ndb-wrap,.nds-wrap,.ndg-wrap){box-sizing:border-box!important;width:min(calc(100% - 2 * var(--nd-content-gutter)),var(--nd-content-width))!important;max-width:var(--nd-content-width)!important;margin-left:auto!important;margin-right:auto!important;padding-left:0!important;padding-right:0!important} html body :is(.ndb,.nds,.ndg){box-sizing:border-box!important;width:100%!important;max-width:100%!important;margin-left:0!important;margin-right:0!important;padding-left:0!important;padding-right:0!important;text-align:left} html body :is(.aui-container,.nd-native-article__content,.nds-article-content,.nd-content-container) :is(.aui-container,.nd-content-container,.ndb-wrap,.nds-wrap,.ndg-wrap,.nds-editorial-wrap){width:100%!important;max-width:100%!important;margin-left:0!important;margin-right:0!important;padding-left:0!important;padding-right:0!important} html body.nds-shared :is(.nd-home-service-grid,.nds-editorial-wrap){width:100%!important;max-width:100%!important;margin-left:0!important;margin-right:0!important} html body.nds-shared :is(.aui-content-copy,.aui-section-heading,.aui-section-note,.nds-legal-copy,.nd-legal-copy):not(.aui-container):not(.nd-content-container){width:100%!important;max-width:none!important;text-align:left!important} html body.nds-shared :is(.aui-section-heading p,.nds-learning-closing h2,.aui-closing-inner>div){max-width:none!important;text-align:left!important} html body :is(.nds-article-section,.ndb-wrap>.ndb-card,.nds-wrap>.nds-card){box-sizing:border-box!important;width:100%!important;max-width:100%!important;padding-left:0!important;padding-right:0!important;text-align:left!important} html body:not(.wp-admin) :is(.nds-article-section,.ndb-card,.nds-card,.aui-content-copy,.nds-legal-copy,.nd-legal-copy,.nd-native-article__content,.nds-article-content,.nds-editorial-page-content,.nd-content-container)>:is(p,h1,h2,h3,h4,ul,ol):not(:has(img,svg,canvas,iframe,video,math,mjx-container,.MathJax)){max-width:none!important;text-align:left!important} html body :is(.ndb,.nds) :is(.nds-comparison-grid,.ndb-comparison-grid,.ndb-figure-grid){width:100%!important;max-width:100%!important;margin-left:0!important;margin-right:0!important} html body.nds-shared main.site-main>.page-content:not(.aui-container):not(:has(.aui-container,.nd-native-article__content,.nds-article-content,.nd-content-container,.nd-home-wrap,.nd-product-wrap,.nd-company-main,.ndlh-wrap,.nd-blog-main,.ndhub-wrap,.ndth-wrap,.ndts-wrap,.ndareas-wrap,.ndarea-wrap,.ndeu-wrap,.nd-contact-wrap,.nd-downloads-wrap,.nd-plain-wrap,.nd-account-hero-inner,.nd-account-main,.ndb-wrap,.nds-wrap,.ndg-wrap)){box-sizing:border-box!important;width:min(calc(100% - 2 * var(--nd-content-gutter)),var(--nd-content-width))!important;max-width:var(--nd-content-width)!important;margin-left:auto!important;margin-right:auto!important;padding-left:0!important;padding-right:0!important} html body :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(figure,.wp-caption,.nd-media-block):not(.wp-block-gallery){ float:none!important;margin-left:auto!important;margin-right:auto!important;max-width:100%;box-sizing:border-box; } html body :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(.nd-content-image,figure>img,figure>a>img,figure>picture>img,.wp-caption>img,.wp-caption>a>img,.nd-media-block>img,.nd-media-block>a>img,.nd-media-block>picture>img){ display:block!important;float:none!important;margin-left:auto!important;margin-right:auto!important;max-width:100%!important;box-sizing:border-box; } html body :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(.nd-media-link,.nd-media-block>picture){display:block;max-width:100%;margin-left:auto!important;margin-right:auto!important} html body :is(.nd-media-group,.ndb-figure-grid,.ndb-selection-grid,.nds-figure-grid,.ndb-comparison-grid,.nds-comparison-grid,.nd-comparison-grid,.wp-block-gallery){ display:flex!important;flex-wrap:wrap!important;justify-content:center!important;align-items:flex-start!important;gap:24px!important; max-width:100%!important;margin-left:auto!important;margin-right:auto!important;box-sizing:border-box; } html body :is(.ndb-figure-grid,.ndb-selection-grid,.nds-figure-grid,.ndb-comparison-grid,.nds-comparison-grid,.nd-comparison-grid)>:is(figure,.ndb-figure-item,.nds-figure-item,.ndb-comparison-item,.nds-comparison-item,.nd-comparison-item,.ndb-interactive-bundle,.nds-interactive-bundle){ flex:0 1 calc((100% - 24px)\/2)!important;width:calc((100% - 24px)\/2)!important;min-width:0!important;max-width:100%!important;margin:0!important;box-sizing:border-box; } html body .nd-media-group>:is(img,a,picture,figure,.nd-media-item){flex:0 1 auto;min-width:0;max-width:calc((100% - 24px)\/2)!important;margin:0!important;box-sizing:border-box} html body .nd-media-group>:is(a,picture,.nd-media-item)>img{max-width:100%!important} html body .nd-media-group>:is(.nd-content-image,.nd-media-link,.nd-media-item){margin-left:0!important;margin-right:0!important} html body :is(.nd-media-group,.ndb-figure-grid,.ndb-selection-grid,.nds-figure-grid,.ndb-comparison-grid,.nds-comparison-grid,.nd-comparison-grid)>:is(br,p:empty){display:none!important} html body .nd-media-helper{display:none!important} @media(max-width:1020px){ html body :is(.ndb-figure-grid,.ndb-selection-grid,.nds-figure-grid,.ndb-comparison-grid,.nds-comparison-grid,.nd-comparison-grid)>:is(figure,.ndb-figure-item,.nds-figure-item,.ndb-comparison-item,.nds-comparison-item,.nd-comparison-item,.ndb-interactive-bundle,.nds-interactive-bundle){flex-basis:100%!important;width:100%!important} } @media(max-width:640px){ html body .nd-media-group>:is(img,a,picture,figure,.nd-media-item){max-width:100%!important} } html body.nds-shared .aui.nds-content-scope :is(h1,h2,h3,h4,h5,h6), html body.nds-shared :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(h1,h2,h3,h4,h5,h6), html body :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(h1,h2,h3,h4,h5,h6){ font-family:Outfit,Roboto,Arial,sans-serif!important;font-weight:600!important; font-style:normal;letter-spacing:-.02em;text-transform:none;overflow-wrap:anywhere; } html body.nds-shared :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(h1,h2,h3,h4,h5,h6)>:is(span,strong,b,a), html body :is(main,.ndb,.nds,.ndg,.nd-content-container) :is(h1,h2,h3,h4,h5,h6)>:is(span,strong,b,a){ font-family:inherit!important;font-weight:inherit!important;font-size:inherit!important;line-height:inherit!important; } html body.nds-shared .aui.nds-content-scope .nd-section-title, html body :is(.nd-section-title,.ndb-card>h2,.nds-card>h2){ display:block!important;box-sizing:border-box;width:100%!important;max-width:none!important; color:#193645!important;font-size:32px!important;font-weight:600!important;line-height:1.25!important; text-align:left!important;letter-spacing:-.02em!important; margin:0 0 20px!important;padding:0 0 12px!important; border:0!important;border-bottom:1px solid #dbe5ec!important;border-radius:0!important; background:none!important;box-shadow:none!important; } html body .nd-section-title>:is(span,strong,b,a){color:inherit!important} html body .nd-section-title::after{content:none!important} html body :is(.nd-heading-section,.nd-heading-step){border-top:0!important} html body .nd-heading-step{display:block!important;padding-left:0!important;padding-right:0!important} html body .nd-heading-step>:is(.ndg-step__body,.nds-step-body){width:100%!important;max-width:none!important;margin-left:0!important;padding-left:0!important} html body :is(.nd-heading-number-source,.nd-heading-divider){display:none!important} html body .nd-section-title>.nd-heading-number{display:inline!important;color:inherit!important;background:none!important;border:0!important;border-radius:0!important;box-shadow:none!important;padding:0!important;margin:0!important;font:inherit!important;white-space:normal} @media(max-width:640px){html body.nds-shared .aui.nds-content-scope .nd-section-title,html body :is(.nd-section-title,.ndb-card>h2,.nds-card>h2){font-size:28px!important}} html body{--nd-photo-tint:rgba(220,238,250,.9);--nd-photo-surface:#dceefa} html body :is(.ndlh-area-card,.nd-photo-card){position:relative;isolation:isolate;overflow:hidden;background-color:var(--nd-photo-surface)!important} html body .nd-photo-card{background-image:var(--nd-photo-bg,none);background-size:cover;background-position:center} html body :is(.ndlh-area-card,.nd-photo-card)::before{ content:\"\"!important;position:absolute!important;inset:0!important;z-index:1!important; background:var(--nd-photo-tint)!important;opacity:1!important;pointer-events:none!important; } html body :is(.ndlh-area-card,.nd-photo-card)>img{ position:absolute!important;inset:0!important;width:100%!important;height:100%!important; max-width:100%!important;object-fit:cover!important;margin:0!important;z-index:0!important; } html body :is(.ndlh-area-card,.nd-photo-card)>:is(.ndlh-area-body,.nds-card-body,.nd-photo-card__body){position:relative!important;z-index:2!important} html body .nd-photo-card__body{padding:24px;color:#193645} html body :is(.ndlh-area-card,.nd-photo-card) :is(.ndlh-area-body,.nds-card-body,.nd-photo-card__body) :is(h2,h3,h4){color:#193645!important} html body :is(.ndlh-area-card,.nd-photo-card) :is(.ndlh-area-body,.nds-card-body,.nd-photo-card__body) :is(p,li){color:#425563!important} html body :is(.ndlh-area-card,.nd-photo-card) :is(.ndlh-area-body,.nds-card-body,.nd-photo-card__body) a:not(:is(.aui-button,.button,.wp-element-button,.wp-block-button__link,[role=\"button\"])){color:#125574!important} html body :is(.ndareas-group-head,.ndhub-card-head,.nd-photo-header){ background-color:var(--nd-photo-surface)!important; background-image:linear-gradient(var(--nd-photo-tint),var(--nd-photo-tint)),var(--nd-photo-bg,var(--ndareas-bg,var(--ndhub-bg,none)))!important; background-size:cover!important;background-position:center!important; color:#193645; }\/* nd-shared-components:end *\/\n<\/style>\n<div class=\"nds\" data-health-profile=\"biomedical-research\">\n<div class=\"nds-wrap\">\n<section class=\"nds-executive\"><h2>Classify dermoscopic images as benign or melanoma<\/h2><p>A compact convolutional network built and trained in Neural Designer 8.0.1 scores all 900 images of the ISIC 2016 lesion-classification training release. On 180 held-out testing images it reaches ROC AUC 0.739. At a threshold fixed on the validation images, it flags 26 of 37 melanomas and leaves 98 of 143 benign lesions unflagged. It is an educational research example, not a diagnostic tool.<\/p><div class=\"nds-kpis\"><div class=\"nds-kpi\"><strong>900<\/strong><span>Dermoscopic images (CC0)<\/span><\/div><div class=\"nds-kpi\"><strong>0.739<\/strong><span>Testing ROC AUC<\/span><\/div><div class=\"nds-kpi\"><strong>70.3%<\/strong><span>Testing sensitivity at threshold 0.11<\/span><\/div><div class=\"nds-kpi\"><strong>68.5%<\/strong><span>Testing specificity at threshold 0.11<\/span><\/div><\/div><div class=\"nds-actions\"><a class=\"aui aui-button\" href=\"#6-clinical-validation\">Review the test results<\/a><a class=\"aui aui-button aui-button--secondary\" href=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma-isic2016-900-20261006.zip\">Download data and model (ZIP)<\/a><\/div><\/section>\n<ul class=\"nds-toc\"><li><a href=\"#1-intended-use\">Question and use<\/a><\/li><li><a href=\"#2-cohort-endpoint\">Images and endpoint<\/a><\/li><li><a href=\"#3-model\">Model<\/a><\/li><li><a href=\"#4-training\">Training<\/a><\/li><li><a href=\"#5-selection\">Selection<\/a><\/li><li><a href=\"#6-clinical-validation\">Validation<\/a><\/li><li><a href=\"#7-workflow\">Workflow<\/a><\/li><li><a href=\"#8-safety\">Safety and validity<\/a><\/li><\/ul>\n<section id=\"1-intended-use\" class=\"nds-card\"><h2>1. Clinical question and intended use<\/h2><p>Dermoscopy magnifies the pigment structures of a skin lesion, and melanoma is the skin cancer that dermoscopic triage most needs to catch. This example asks a narrow research question: how well can a small convolutional network, trained from scratch on 540 labeled dermoscopic images, rank new images toward the melanoma label?<\/p><p>The example shows the complete Neural Designer workflow for image classification: importing an image folder, training a network, choosing an operating threshold on validation data and reporting every error on testing data. It does not establish clinical performance.<\/p><div class=\"nds-value-grid\"><div class=\"nds-value\"><h3>Licensed, traceable images<\/h3><p>Every image comes from the CC0 ISIC 2016 release, with source and file hashes in the manifest.<\/p><\/div><div class=\"nds-value\"><h3>Threshold fixed in advance<\/h3><p>The operating point is chosen on validation images before the testing images are evaluated.<\/p><\/div><div class=\"nds-value\"><h3>Complete error accounting<\/h3><p>Report missed melanomas and false alarms, not only accuracy.<\/p><\/div><\/div><div class=\"nds-audience\"><span>Dermatology research<\/span><span>Biomedical imaging<\/span><span>Model validation<\/span><span>Medical ML education<\/span><\/div><div class=\"nds-note nds-use-boundary\"><strong>Intended-use boundary.<\/strong> Educational research example only. The model is not clinically validated and must not be used to diagnose skin lesions. A qualified clinician must assess any suspicious lesion.<\/div><\/section>\n<section id=\"2-cohort-endpoint\" class=\"nds-card\"><h2>2. Images, labels and endpoint<\/h2><p>The data are the 900 dermoscopic images of the <a href=\"https:\/\/challenge.isic-archive.com\/data\/#2016\">ISIC 2016 Challenge, Part 3 training release<\/a>, published by the International Skin Imaging Collaboration under CC0 1.0. Each image has one label: 727 are benign and 173 are malignant (melanoma), a melanoma share of 19.2%. The endpoint is that label; the positive class is malignant.<\/p><p>The original JPEG images range from 576 \u00d7 542 to 4288 \u00d7 2848 pixels (median 1024 \u00d7 768). Each one is converted to RGB and resized to 300 \u00d7 300 pixels with Lanczos resampling, without cropping, so non-square lesions are slightly distorted. This is the same preparation as the 102-image OpenNN example, whose images are bit-identical to the corresponding files here.<\/p><figure class=\"nds-centered-figure\"><img loading=\"lazy\" decoding=\"async\" src=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg\" alt=\"Eight dermoscopic training images: four benign lesions in the top row and four melanomas in the bottom row\" width=\"996\" height=\"492\" loading=\"lazy\"><figcaption>Training images after preparation. Top row: benign labels (ISIC_0000000, 0000001, 0000006, 0000008). Bottom row: malignant labels (ISIC_0000026, 0000030, 0000031, 0000049).<\/figcaption><\/figure><p>Neural Designer imports the folder with one subfolder per class and assigns images at random to training (60%), validation (20%) and testing (20%) subsets. The project stores this assignment.<\/p><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Subset<\/th><th scope=\"col\">Images<\/th><th scope=\"col\">Benign<\/th><th scope=\"col\">Malignant<\/th><th scope=\"col\">Melanoma share<\/th><th scope=\"col\">Purpose<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">Training<\/th><td>540<\/td><td>438<\/td><td>102<\/td><td>18.9%<\/td><td>Fit the network parameters<\/td><\/tr><tr><th scope=\"row\">Validation<\/th><td>180<\/td><td>146<\/td><td>34<\/td><td>18.9%<\/td><td>Compare candidates and fix the threshold<\/td><\/tr><tr><th scope=\"row\">Testing<\/th><td>180<\/td><td>143<\/td><td>37<\/td><td>20.6%<\/td><td>Final evaluation, used once<\/td><\/tr><tr><th scope=\"row\">Total<\/th><td>900<\/td><td>727<\/td><td>173<\/td><td>19.2%<\/td><td>ISIC 2016 Part 3 training release<\/td><\/tr><\/tbody><\/table><\/div><div class=\"nds-note nds-note--provenance\"><strong>Provenance.<\/strong> The split is by image. The ISIC 2016 release does not include patient identifiers, so lesions from the same patient may fall in different subsets. The official ISIC 2016 test release (379 images) is not used here. The download includes the manifest with the SHA-256 of every source JPEG and prepared BMP.<\/div><\/section>\n<section id=\"3-model\" class=\"nds-card\"><h2>3. Model<\/h2><p>The network is the default image-classification architecture that Neural Designer 8.0.1 creates for 300 \u00d7 300 RGB images: one convolutional layer with eight 3 \u00d7 3 filters, max pooling and a single sigmoid output. It has 180,225 trainable parameters.<\/p><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Layer<\/th><th scope=\"col\">Input shape<\/th><th scope=\"col\">Output shape<\/th><th scope=\"col\">Configuration<\/th><th scope=\"col\">Parameters<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">Scaling<\/th><td>300 \u00d7 300 \u00d7 3<\/td><td>300 \u00d7 300 \u00d7 3<\/td><td>ImageMinMax pixel scaling<\/td><td>0<\/td><\/tr><tr><th scope=\"row\">Convolutional<\/th><td>300 \u00d7 300 \u00d7 3<\/td><td>300 \u00d7 300 \u00d7 8<\/td><td>8 filters, 3 \u00d7 3, stride 1, same padding, ReLU<\/td><td>224<\/td><\/tr><tr><th scope=\"row\">Pooling<\/th><td>300 \u00d7 300 \u00d7 8<\/td><td>150 \u00d7 150 \u00d7 8<\/td><td>Max pooling, 2 \u00d7 2, stride 2<\/td><td>0<\/td><\/tr><tr><th scope=\"row\">Flatten<\/th><td>150 \u00d7 150 \u00d7 8<\/td><td>180,000<\/td><td><\/td><td>0<\/td><\/tr><tr><th scope=\"row\">Dense (classification)<\/th><td>180,000<\/td><td>1<\/td><td>Sigmoid<\/td><td>180,001<\/td><\/tr><\/tbody><\/table><\/div><p><strong>Output contract.<\/strong> The output is a score between 0 and 1 for the malignant class; the benign score is its complement. The scores are not calibrated probabilities. The image is flagged as suspicious when the malignant score reaches the decision threshold described in Section 5.<\/p><figure class=\"nds-architecture-figure\"><a class=\"nds-architecture-link\" href=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-architecture-20261006.png\" target=\"_blank\" rel=\"noopener\"><img loading=\"lazy\" decoding=\"async\" class=\"nds-architecture\" src=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-architecture-20261006.png\" alt=\"Network architecture: scaling, convolutional layer with eight filters, max pooling, flatten layer with 180,000 values and one sigmoid output named benign_malignant\" width=\"2320\" height=\"1096\" loading=\"lazy\"><\/a><figcaption>Network architecture. Open the image to inspect the layer dimensions.<\/figcaption><\/figure><\/section>\n<section id=\"4-training\" class=\"nds-card\"><h2>4. Training strategy<\/h2><p>Training minimizes the cross-entropy error with the adaptive moment estimation (Adam) optimizer: learning rate 0.0001, mini-batches of 32 images, L2 regularization with weight 0.01 and 100 epochs. Training ran on an NVIDIA GeForce RTX 5070 Ti GPU.<\/p><p>The training and validation errors below are the mean binary cross-entropy of the malignant score over the images, without the regularization term. For reference, a constant score of 0.5 gives 0.693.<\/p><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Adaptive moment estimation results<\/th><th scope=\"col\">Value<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">Epochs number<\/th><td>100<\/td><\/tr><tr><th scope=\"row\">Elapsed time<\/th><td>00:00:29<\/td><\/tr><tr><th scope=\"row\">Stopping criterion<\/th><td>Maximum epochs number<\/td><\/tr><tr><th scope=\"row\">Training error<\/th><td>0.261<\/td><\/tr><tr><th scope=\"row\">Validation error<\/th><td>0.451<\/td><\/tr><\/tbody><\/table><\/div><div class=\"nd-comparison-grid\"><div class=\"nd-comparison-item\"><figure class=\"nd-media-block\"><div class=\"nd-chart-bundle\" data-native-chart=\"nd-melanoma900-history-20261006\" role=\"group\" aria-label=\"Interactive chart; scroll horizontally on narrow screens\" tabindex=\"0\"><div class=\"nd-chart-host\" id=\"nd-melanoma900-history-20261006\" role=\"img\" aria-label=\"Training and validation cross-entropy error by epoch\"><\/div><p class=\"nd-chart-fallback\">Training and validation cross-entropy error by epoch. Enable JavaScript to explore the chart.<\/p><script type=\"application\/json\">{\"config\":{\"displayModeBar\":true,\"displaylogo\":false,\"editable\":false,\"responsive\":true,\"scrollZoom\":true,\"toImageButtonOptions\":{\"filename\":\"neural-designer-chart\",\"format\":\"svg\"}},\"data\":[{\"connectgaps\":false,\"line\":{\"color\":\"#529cc2\",\"dash\":\"solid\",\"width\":2},\"mode\":\"lines\",\"name\":\"Training error\",\"showlegend\":true,\"type\":\"scatter\",\"x\":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100],\"y\":[0.753,0.516,0.472,0.464,0.472,0.463,0.472,0.481,0.476,0.465,0.452,0.462,0.438,0.438,0.433,0.447,0.435,0.43,0.426,0.431,0.432,0.435,0.422,0.426,0.417,0.413,0.408,0.405,0.393,0.378,0.389,0.379,0.384,0.375,0.371,0.365,0.364,0.359,0.352,0.358,0.365,0.359,0.351,0.363,0.346,0.35,0.341,0.341,0.353,0.35,0.338,0.354,0.354,0.353,0.351,0.341,0.339,0.321,0.325,0.322,0.329,0.324,0.321,0.324,0.31,0.314,0.318,0.311,0.307,0.3,0.298,0.298,0.296,0.296,0.296,0.292,0.304,0.298,0.289,0.286,0.288,0.284,0.289,0.286,0.274,0.275,0.274,0.273,0.278,0.28,0.266,0.274,0.266,0.275,0.265,0.277,0.275,0.27,0.26,0.255,0.261]},{\"connectgaps\":false,\"line\":{\"color\":\"#ed982a\",\"dash\":\"solid\",\"width\":2},\"mode\":\"lines\",\"name\":\"Validation error\",\"showlegend\":true,\"type\":\"scatter\",\"x\":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100],\"y\":[0.46,0.504,0.488,0.48,0.47,0.468,0.499,0.502,0.451,0.446,0.494,0.438,0.45,0.458,0.47,0.456,0.416,0.431,0.445,0.431,0.445,0.434,0.445,0.435,0.434,0.433,0.451,0.461,0.434,0.456,0.434,0.433,0.462,0.475,0.454,0.449,0.473,0.472,0.432,0.441,0.415,0.436,0.431,0.428,0.447,0.444,0.409,0.436,0.426,0.409,0.466,0.44,0.406,0.487,0.465,0.466,0.43,0.431,0.465,0.418,0.462,0.459,0.497,0.426,0.437,0.444,0.474,0.5,0.454,0.471,0.458,0.454,0.449,0.405,0.438,0.481,0.426,0.439,0.417,0.445,0.448,0.488,0.467,0.443,0.455,0.399,0.449,0.471,0.435,0.428,0.436,0.425,0.406,0.416,0.44,0.448,0.459,0.396,0.451,0.411,0.451]}],\"layout\":{\"autosize\":true,\"font\":{\"color\":\"#30343b\",\"family\":\"Arial, sans-serif\",\"size\":14},\"hovermode\":\"closest\",\"legend\":{\"orientation\":\"h\",\"x\":0.5,\"xanchor\":\"center\",\"y\":-0.24,\"yanchor\":\"top\"},\"margin\":{\"l\":95,\"r\":45,\"t\":65,\"b\":95},\"paper_bgcolor\":\"#fff\",\"plot_bgcolor\":\"#fff\",\"title\":{\"automargin\":true,\"pad\":{\"t\":48},\"text\":\"Training history\",\"x\":0.5,\"xanchor\":\"center\",\"y\":1,\"yanchor\":\"top\",\"yref\":\"container\",\"font\":{\"size\":17}},\"xaxis\":{\"automargin\":true,\"gridcolor\":\"#d9dde3\",\"linecolor\":\"#9aa3ad\",\"range\":[0,100],\"title\":{\"text\":\"Epoch\"},\"zerolinecolor\":\"#aeb6c2\"},\"yaxis\":{\"automargin\":true,\"gridcolor\":\"#d9dde3\",\"linecolor\":\"#9aa3ad\",\"range\":[0,0.8],\"title\":{\"text\":\"Cross-entropy error\"},\"zerolinecolor\":\"#aeb6c2\"},\"height\":500}}<\/script><\/div><figcaption>Training and validation error at each epoch, as reported by Neural Designer.<\/figcaption><\/figure><\/div><\/div><p>The training error falls from 0.753 to 0.261, while the validation error stays between about 0.40 and 0.50 after the first epochs. The network keeps fitting the training images without improving on unseen ones: with 540 training images, the data limit what this architecture can learn. The saved model is the one at epoch 100.<\/p><\/section>\n<section id=\"5-selection\" class=\"nds-card\"><h2>5. Model selection and threshold<\/h2><p>Five configurations were trained on the same split and compared by their final validation error. The testing images were not used for this choice.<\/p><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Candidate<\/th><th scope=\"col\">Configuration<\/th><th scope=\"col\">Final validation error<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">No regularization<\/th><td>Default network, Adam, learning rate 0.0001, 100 epochs<\/td><td>0.482<\/td><\/tr><tr><th scope=\"row\"><strong>L2 regularization 0.01 (selected)<\/strong><\/th><td>Same, with L2 weight 0.01<\/td><td><strong>0.451<\/strong><\/td><\/tr><tr><th scope=\"row\">Higher learning rate<\/th><td>No regularization, learning rate 0.001<\/td><td>0.476<\/td><\/tr><tr><th scope=\"row\">Shorter training<\/th><td>No regularization, 50 epochs<\/td><td>0.464<\/td><\/tr><tr><th scope=\"row\">Hidden dense layer<\/th><td>128 ReLU neurons after flatten (23.0 million parameters)<\/td><td>2.989 (training failed)<\/td><\/tr><\/tbody><\/table><\/div><p>The first four candidates differ by at most 0.031, which is within the epoch-to-epoch fluctuation of the validation error (up to 0.08), so the choice among them is weak; the selected model keeps the regularization. The larger network with a hidden layer of 128 neurons did not train: its error stayed near 3 and its malignant score saturated at 0 for every image checked.<\/p><h3>Decision threshold<\/h3><p>Melanomas are 19% of the training images, so the model&#8217;s scores are low: the median malignant score is 0.075 for benign and 0.171 for malignant testing images. The default threshold of 0.5 therefore flags very few images. The threshold was set on the validation images, as the point of the validation ROC curve closest to perfect classification: <strong>0.11<\/strong>, with validation ROC AUC 0.750, sensitivity 76.5% and specificity 65.8%.<\/p><p><strong>Baseline.<\/strong> Predicting \u201cbenign\u201d for every image is correct for 143 of the 180 testing images (79.4% accuracy) but detects no melanoma. Accuracy alone is therefore not a useful measure here.<\/p><\/section>\n<section id=\"6-clinical-validation\" class=\"nds-card\"><h2>6. Clinical validation<\/h2><p>The final model is evaluated once on the 180 testing images: 37 malignant (20.6%) and 143 benign. Neural Designer reports a testing ROC AUC of 0.739. Calculated from the same scores, the 95% confidence interval is about 0.64\u20130.84 (Hanley\u2013McNeil).<\/p><div class=\"nd-comparison-grid\"><div class=\"nd-comparison-item\"><figure class=\"nd-media-block\"><div class=\"nd-chart-bundle\" data-native-chart=\"nd-melanoma900-roc-20261006\" role=\"group\" aria-label=\"Interactive chart; scroll horizontally on narrow screens\" tabindex=\"0\"><div class=\"nd-chart-host\" id=\"nd-melanoma900-roc-20261006\" role=\"img\" aria-label=\"ROC curve of the testing images with the chosen threshold marked\"><\/div><p class=\"nd-chart-fallback\">ROC curve of the testing images with the chosen threshold marked. Enable JavaScript to explore the chart.<\/p><script type=\"application\/json\">{\"config\":{\"displayModeBar\":true,\"displaylogo\":false,\"editable\":false,\"responsive\":true,\"scrollZoom\":true,\"toImageButtonOptions\":{\"filename\":\"neural-designer-chart\",\"format\":\"svg\"}},\"data\":[{\"line\":{\"color\":\"#9aa3ad\",\"dash\":\"dash\",\"width\":1.5},\"mode\":\"lines\",\"name\":\"Random classifier\",\"type\":\"scatter\",\"x\":[0,1],\"y\":[0,1],\"hoverinfo\":\"skip\"},{\"line\":{\"color\":\"#529cc2\",\"width\":2.5,\"shape\":\"hv\"},\"mode\":\"lines\",\"name\":\"ROC curve (AUC 0.739)\",\"type\":\"scatter\",\"x\":[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.007,0.014,0.028,0.028,0.028,0.028,0.028,0.028,0.028,0.028,0.028,0.028,0.028,0.035,0.035,0.035,0.035,0.035,0.035,0.042,0.049,0.049,0.049,0.049,0.049,0.049,0.049,0.049,0.049,0.056,0.056,0.063,0.07,0.077,0.077,0.098,0.112,0.147,0.147,0.161,0.196,0.21,0.252,0.287,0.315,0.364,0.413,0.469,0.559,0.615,0.657,0.685,0.72,0.811,0.888,1.0,1.0],\"y\":[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.027,0.054,0.054,0.054,0.081,0.081,0.081,0.081,0.081,0.081,0.081,0.081,0.081,0.081,0.081,0.108,0.108,0.108,0.108,0.108,0.108,0.108,0.108,0.108,0.108,0.135,0.162,0.162,0.162,0.162,0.189,0.189,0.189,0.189,0.189,0.189,0.216,0.243,0.243,0.27,0.324,0.351,0.405,0.459,0.459,0.459,0.459,0.459,0.514,0.568,0.568,0.622,0.649,0.649,0.703,0.703,0.73,0.73,0.784,0.838,0.838,0.865,0.919,0.973,1.0,1.0,1.0],\"text\":[\"\",\"Threshold 1.00\",\"Threshold 0.99\",\"Threshold 0.98\",\"Threshold 0.97\",\"Threshold 0.96\",\"Threshold 0.95\",\"Threshold 0.94\",\"Threshold 0.93\",\"Threshold 0.92\",\"Threshold 0.91\",\"Threshold 0.90\",\"Threshold 0.89\",\"Threshold 0.88\",\"Threshold 0.87\",\"Threshold 0.86\",\"Threshold 0.85\",\"Threshold 0.84\",\"Threshold 0.83\",\"Threshold 0.82\",\"Threshold 0.81\",\"Threshold 0.80\",\"Threshold 0.79\",\"Threshold 0.78\",\"Threshold 0.77\",\"Threshold 0.76\",\"Threshold 0.75\",\"Threshold 0.74\",\"Threshold 0.73\",\"Threshold 0.72\",\"Threshold 0.71\",\"Threshold 0.70\",\"Threshold 0.69\",\"Threshold 0.68\",\"Threshold 0.67\",\"Threshold 0.66\",\"Threshold 0.65\",\"Threshold 0.64\",\"Threshold 0.63\",\"Threshold 0.62\",\"Threshold 0.61\",\"Threshold 0.60\",\"Threshold 0.59\",\"Threshold 0.58\",\"Threshold 0.57\",\"Threshold 0.56\",\"Threshold 0.55\",\"Threshold 0.54\",\"Threshold 0.53\",\"Threshold 0.52\",\"Threshold 0.51\",\"Threshold 0.50\",\"Threshold 0.49\",\"Threshold 0.48\",\"Threshold 0.47\",\"Threshold 0.46\",\"Threshold 0.45\",\"Threshold 0.44\",\"Threshold 0.43\",\"Threshold 0.42\",\"Threshold 0.41\",\"Threshold 0.40\",\"Threshold 0.39\",\"Threshold 0.38\",\"Threshold 0.37\",\"Threshold 0.36\",\"Threshold 0.35\",\"Threshold 0.34\",\"Threshold 0.33\",\"Threshold 0.32\",\"Threshold 0.31\",\"Threshold 0.30\",\"Threshold 0.29\",\"Threshold 0.28\",\"Threshold 0.27\",\"Threshold 0.26\",\"Threshold 0.25\",\"Threshold 0.24\",\"Threshold 0.23\",\"Threshold 0.22\",\"Threshold 0.21\",\"Threshold 0.20\",\"Threshold 0.19\",\"Threshold 0.18\",\"Threshold 0.17\",\"Threshold 0.16\",\"Threshold 0.15\",\"Threshold 0.14\",\"Threshold 0.13\",\"Threshold 0.12\",\"Threshold 0.11\",\"Threshold 0.10\",\"Threshold 0.09\",\"Threshold 0.08\",\"Threshold 0.07\",\"Threshold 0.06\",\"Threshold 0.05\",\"Threshold 0.04\",\"Threshold 0.03\",\"Threshold 0.02\",\"Threshold 0.01\",\"Threshold 0.00\",\"\"],\"hovertemplate\":\"%{text}<br>False positive rate %{x:.3f}<br>Sensitivity %{y:.3f}<extra><\\\/extra>\"},{\"marker\":{\"color\":\"#ed982a\",\"size\":12,\"line\":{\"color\":\"#30343b\",\"width\":1}},\"mode\":\"markers\",\"name\":\"Threshold 0.11 (chosen on validation)\",\"type\":\"scatter\",\"x\":[0.315],\"y\":[0.703],\"hovertemplate\":\"Threshold 0.11<br>False positive rate %{x:.3f}<br>Sensitivity %{y:.3f}<extra><\\\/extra>\"}],\"layout\":{\"autosize\":true,\"font\":{\"color\":\"#30343b\",\"family\":\"Arial, sans-serif\",\"size\":14},\"hovermode\":\"closest\",\"legend\":{\"orientation\":\"h\",\"x\":0.5,\"xanchor\":\"center\",\"y\":-0.24,\"yanchor\":\"top\"},\"margin\":{\"l\":95,\"r\":45,\"t\":65,\"b\":95},\"paper_bgcolor\":\"#fff\",\"plot_bgcolor\":\"#fff\",\"title\":{\"automargin\":true,\"pad\":{\"t\":48},\"text\":\"ROC curve on the testing images\",\"x\":0.5,\"xanchor\":\"center\",\"y\":1,\"yanchor\":\"top\",\"yref\":\"container\",\"font\":{\"size\":17}},\"xaxis\":{\"automargin\":true,\"gridcolor\":\"#d9dde3\",\"linecolor\":\"#9aa3ad\",\"range\":[0,1],\"title\":{\"text\":\"False positive rate (1 \\u2212 specificity)\"},\"zerolinecolor\":\"#aeb6c2\"},\"yaxis\":{\"automargin\":true,\"gridcolor\":\"#d9dde3\",\"linecolor\":\"#9aa3ad\",\"range\":[0,1],\"title\":{\"text\":\"Sensitivity\"},\"zerolinecolor\":\"#aeb6c2\"},\"height\":560}}<\/script><\/div><figcaption>ROC curve on the testing images, from Neural Designer&#8217;s sensitivity and specificity table. The orange point is the threshold fixed on the validation images.<\/figcaption><\/figure><\/div><\/div><h3>Operating point at threshold 0.11<\/h3><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Actual \/ predicted at 0.11<\/th><th scope=\"col\">Flagged (malignant)<\/th><th scope=\"col\">Not flagged (benign)<\/th><th scope=\"col\">Total<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">Malignant label<\/th><td>26<\/td><td>11<\/td><td>37<\/td><\/tr><tr><th scope=\"row\">Benign label<\/th><td>45<\/td><td>98<\/td><td>143<\/td><\/tr><tr><th scope=\"row\">Total<\/th><td>71<\/td><td>109<\/td><td>180<\/td><\/tr><\/tbody><\/table><\/div><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Testing metric<\/th><th scope=\"col\">Threshold 0.11<\/th><th scope=\"col\">Count-based reading<\/th><th scope=\"col\">Threshold 0.5<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">Sensitivity<\/th><td>70.3%<\/td><td>26 of 37 melanomas flagged<\/td><td>10.8% (4 of 37)<\/td><\/tr><tr><th scope=\"row\">Specificity<\/th><td>68.5%<\/td><td>98 of 143 benign lesions not flagged<\/td><td>97.2% (139 of 143)<\/td><\/tr><tr><th scope=\"row\">Precision \/ observed PPV<\/th><td>36.6%<\/td><td>26 of 71 flagged images are melanomas<\/td><td>50.0% (4 of 8)<\/td><\/tr><tr><th scope=\"row\">Observed NPV<\/th><td>89.9%<\/td><td>98 of 109 unflagged images are benign<\/td><td>80.8% (139 of 172)<\/td><\/tr><tr><th scope=\"row\">Accuracy<\/th><td>68.9%<\/td><td>124 of 180 images classified correctly<\/td><td>79.4% (143 of 180)<\/td><\/tr><tr><th scope=\"row\">F1 score<\/th><td>0.481<\/td><td>Harmonic mean of precision and sensitivity<\/td><td>0.178<\/td><\/tr><\/tbody><\/table><\/div><p>At 0.5 the model behaves almost like the all-benign baseline: same accuracy, and only 4 of 37 melanomas detected. At 0.11 it detects 26 of 37 melanomas, at the cost of 45 false alarms among 143 benign lesions. On the testing images, the ROC point closest to perfect classification lies at a threshold of 0.13. That value is descriptive only, because it was found on the testing data.<\/p><div class=\"nds-note\"><strong>Clinical interpretation.<\/strong> The model learns a real but modest signal from the images. Eleven of the 37 testing melanomas score below the threshold, and only 36.6% of flagged images are melanomas at this 20.6% prevalence. On the training images the ROC AUC is 0.955, so performance on new images is much lower than on the images it was fitted to.<\/div><\/section>\n<section id=\"7-workflow\" class=\"nds-card\"><h2>7. Workflow and reproducibility<\/h2><p>A responsible real-world analogue would begin after a lesion has entered an approved clinical pathway. It would check image quality and format, calculate a versioned score, and route flagged and unflagged lesions alike to clinician review, with histopathology as the reference where indicated.<\/p><div class=\"nd-output-examples\" id=\"calculate-outputs\"><h3>Calculating outputs for new images<\/h3><p>Neural Designer&#8217;s <em>Calculate outputs<\/em> task scores one image at a time. The two testing images below show one melanoma above the threshold and one benign lesion below it.<\/p><div class=\"nd-comparison-grid\"><div class=\"nd-comparison-item nd-output-example\" data-output-example=\"1\"><h4>Example 1<\/h4><figure class=\"nd-media-block\"><img loading=\"lazy\" decoding=\"async\" src=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-output-isic_0000004-20261006.jpg\" alt=\"Dermoscopic image ISIC_0000004 (malignant label, testing subset)\" width=\"300\" height=\"300\" loading=\"lazy\"><figcaption>Input image: <code>malignant\/ISIC_0000004.bmp<\/code> (testing subset, malignant label)<\/figcaption><\/figure><p class=\"nd-output-result\"><strong>Malignant score: 0.8151<\/strong><br>Flagged: the score is above the threshold of 0.11.<\/p><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Class<\/th><th scope=\"col\">Output score<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">benign<\/th><td>0.1849<\/td><\/tr><tr><th scope=\"row\">malignant<\/th><td>0.8151<\/td><\/tr><\/tbody><\/table><\/div><\/div><div class=\"nd-comparison-item nd-output-example\" data-output-example=\"2\"><h4>Example 2<\/h4><figure class=\"nd-media-block\"><img loading=\"lazy\" decoding=\"async\" src=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-output-isic_0000028-20261006.jpg\" alt=\"Dermoscopic image ISIC_0000028 (benign label, testing subset)\" width=\"300\" height=\"300\" loading=\"lazy\"><figcaption>Input image: <code>benign\/ISIC_0000028.bmp<\/code> (testing subset, benign label)<\/figcaption><\/figure><p class=\"nd-output-result\"><strong>Malignant score: 0.0209<\/strong><br>Not flagged: the score is below the threshold of 0.11.<\/p><div class=\"nd-table-scroll\" tabindex=\"0\"><table><thead><tr><th scope=\"col\">Class<\/th><th scope=\"col\">Output score<\/th><\/tr><\/thead><tbody><tr><th scope=\"row\">benign<\/th><td>0.9791<\/td><\/tr><tr><th scope=\"row\">malignant<\/th><td>0.0209<\/td><\/tr><\/tbody><\/table><\/div><\/div><\/div><p>These are illustrative outputs for images already counted in the testing metrics. The scores are uncalibrated.<\/p><\/div><h3>Reproduce the example<\/h3><p>The ZIP contains the 900 prepared images, the manifest, the trained Neural Designer 8.0.1 model (<code>model\/melanoma.ndm<\/code> and its parameter file), the Neural Designer inference script and <code>scores.csv<\/code> with the malignant score of every image and its subset. All the metrics above can be recalculated from <code>scores.csv<\/code>. To score a new image prepared in the same way, run:<\/p><pre><code>model\\melanoma_predict.bat lesion.bmp --table<\/code><\/pre><p>The script reports its predicted class at 0.5; apply the 0.11 threshold to the malignant score instead. To retrain, create an image-classification project in Neural Designer from the <code>melanoma_dataset_bmp<\/code> folder, keep the default network and use the training settings in Section 4. Neural Designer draws a new random split and initialization, so the results will vary; the saved split is listed in <code>model\/melanoma.ndm<\/code> and <code>scores.csv<\/code>.<\/p><div class=\"nds-downloads\"><a class=\"aui aui-button\" href=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma-isic2016-900-20261006.zip\">Download data and model (ZIP, 175 MB)<\/a><a class=\"aui aui-button aui-button--secondary\" href=\"https:\/\/www.neuraldesigner.com\/downloads\/\">Download Neural Designer<\/a><\/div><\/section>\n<section id=\"8-safety\" class=\"nds-card\"><h2>8. Safety, generalizability and governance<\/h2><ul><li><strong>Internal random split only.<\/strong> Training, validation and testing images come from the same release. No external hospital, device, time period or prospective cohort is evaluated.<\/li><li><strong>Patient grouping cannot be verified.<\/strong> The release has no patient identifiers, so related lesions may appear in more than one subset, which can make internal results optimistic.<\/li><li><strong>Few melanomas in the test.<\/strong> With 37 testing melanomas, each missed case changes sensitivity by 2.7 percentage points, and the confidence interval of the AUC is wide.<\/li><li><strong>Challenge selection, not a clinical population.<\/strong> The 19% melanoma share reflects how the challenge images were collected. Predictive values change with prevalence.<\/li><li><strong>Image preparation.<\/strong> Downsampling to 300 \u00d7 300 discards fine dermoscopic structure, the resize distorts non-square images, and no color or artifact normalization is applied.<\/li><li><strong>Uncalibrated scores.<\/strong> A threshold of 0.11 shows that the scores are not probabilities of melanoma.<\/li><li><strong>No subgroup evidence.<\/strong> Skin type, age, sex, anatomical site and acquisition device are not available in the release, so performance across them is unknown.<\/li><li><strong>No clinical-utility evaluation.<\/strong> The example does not compare with dermatologists or measure the effect on patient outcomes.<\/li><\/ul><div class=\"nds-note nds-note--warning\"><strong>Decision boundary.<\/strong> Use this model for education and method research only. Any clinical use would require larger and external data, calibration, a prespecified threshold, regulatory review and evaluation within the clinical pathway.<\/div><\/section>\n<section id=\"references\" class=\"nds-card\"><h2>References<\/h2><ul><li><a href=\"https:\/\/challenge.isic-archive.com\/data\/#2016\">ISIC 2016 Challenge, Part 3 training release<\/a>. International Skin Imaging Collaboration (ISIC). CC0 1.0 Universal.<\/li><li>Gutman D, Codella NCF, Celebi E, Helba B, Marchetti M, Mishra N, Halpern A. <a href=\"https:\/\/arxiv.org\/abs\/1605.01397\">Skin Lesion Analysis toward Melanoma Detection: A Challenge at the International Symposium on Biomedical Imaging (ISBI) 2016, hosted by the International Skin Imaging Collaboration (ISIC)<\/a>. arXiv:1605.01397, 2016.<\/li><li><a href=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma-isic2016-900-20261006.zip\">Prepared images, trained Neural Designer 8.0.1 model and scores<\/a>, 6 October 2026 (ZIP).<\/li><li><a href=\"https:\/\/www.neuraldesigner.com\/learning\/examples\/examples\/\">More Neural Designer examples<\/a><\/li><\/ul><\/section>\n<\/div>\n<\/div>\n<style>\n.nd-chart-bundle{width:100%;max-width:760px;min-width:0;margin:24px auto;box-sizing:border-box;position:relative;overflow-x:auto;text-align:center;background:#fff;color:#30343b}\n.nd-chart-bundle .nd-chart-host{width:100%;height:440px;min-width:0;box-sizing:border-box}\n.nd-chart-bundle:not(.is-ready) .nd-chart-host{position:absolute;visibility:hidden}\n.nd-chart-bundle.is-ready .nd-chart-fallback{display:none!important}\n.nd-chart-bundle .nd-chart-fallback{display:block;margin:0 auto!important;width:100%;height:auto;max-width:100%}\n.nd-chart-bundle>script,.nd-chart-bundle>br,.nd-chart-bundle>p:empty{display:none!important}\n@media(max-width:600px){.nd-chart-bundle{margin:20px auto}.nd-chart-bundle .nd-chart-host{height:440px}}\n.nd-chart-bundle .nd-chart-host{min-width:0!important}.nd-chart-bundle{max-width:100%}html body .nd-comparison-grid>.nd-comparison-item:only-child{flex-basis:min(760px,100%)!important;width:min(760px,100%)!important}.ndb-card details,.nds-card details{margin:20px 0}.ndb-card summary,.nds-card summary{cursor:pointer;font-weight:600}<\/style><script data-noptimize=\"1\" data-cfasync=\"false\" src=\"https:\/\/cdn.plot.ly\/plotly-basic-4.0.0.min.js\"><\/script><script data-noptimize=\"1\">(()=>{const start=()=>{if(!window.Plotly)return;document.querySelectorAll('.nd-chart-bundle').forEach(bundle=>{if(bundle.dataset.started)return;bundle.dataset.started='1';const host=bundle.querySelector('.nd-chart-host'),source=bundle.querySelector('script[type=\"application\/json\"]');if(!host||!source)return;const figure=JSON.parse(source.textContent);window.Plotly.newPlot(host,figure.data,figure.layout,figure.config).then(async()=>{const title=host.querySelector('.gtitle');if(title)host.style.minWidth=Math.ceil(title.getComputedTextLength()+48)+'px';await window.Plotly.Plots.resize(host);bundle.classList.add('is-ready');if(window.ResizeObserver){let timer;new ResizeObserver(()=>{clearTimeout(timer);timer=setTimeout(()=>window.Plotly.Plots.resize(host),80);}).observe(bundle);}}).catch(error=>{bundle.dataset.error=String(error);console.error(error);});});};const ready=()=>{if(window.Plotly){start();return;}let count=0;const timer=setInterval(()=>{if(window.Plotly){clearInterval(timer);start();}else if(++count>200)clearInterval(timer);},50);};if(document.readyState==='loading')document.addEventListener('DOMContentLoaded',ready,{once:true});else ready();})();<\/script>","protected":false},"author":152,"featured_media":0,"template":"","categories":[29],"tags":[],"class_list":["post-24127","learning","type-learning","status-publish","hentry","category-examples"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.4 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Melanoma classification from dermoscopic images - Neural Designer<\/title>\n<meta name=\"description\" content=\"Classify 900 ISIC 2016 dermoscopic images as benign or melanoma with a convolutional network in Neural Designer, from data preparation to testing analysis.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Melanoma classification from dermoscopic images - Neural Designer\" \/>\n<meta property=\"og:description\" content=\"Classify 900 ISIC 2016 dermoscopic images as benign or melanoma with a convolutional network in Neural Designer, from data preparation to testing analysis.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/\" \/>\n<meta property=\"og:site_name\" content=\"Neural Designer\" \/>\n<meta property=\"article:modified_time\" content=\"2026-10-06T17:13:22+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:site\" content=\"@NeuralDesigner\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"9 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/\",\"url\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/\",\"name\":\"Melanoma classification from dermoscopic images - Neural Designer\",\"isPartOf\":{\"@id\":\"https:\/\/www.neuraldesigner.com\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg\",\"datePublished\":\"2026-09-24T11:19:09+00:00\",\"dateModified\":\"2026-10-06T17:13:22+00:00\",\"description\":\"Classify 900 ISIC 2016 dermoscopic images as benign or melanoma with a convolutional network in Neural Designer, from data preparation to testing analysis.\",\"breadcrumb\":{\"@id\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#primaryimage\",\"url\":\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg\",\"contentUrl\":\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg\",\"width\":996,\"height\":492},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.neuraldesigner.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Learning\",\"item\":\"https:\/\/www.neuraldesigner.com\/learning\/\"},{\"@type\":\"ListItem\",\"position\":3,\"name\":\"Melanoma classification from dermoscopic images\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.neuraldesigner.com\/#website\",\"url\":\"https:\/\/www.neuraldesigner.com\/\",\"name\":\"Neural Designer\",\"description\":\"No-code AI Platform\",\"publisher\":{\"@id\":\"https:\/\/www.neuraldesigner.com\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.neuraldesigner.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/www.neuraldesigner.com\/#organization\",\"name\":\"Neural Designer\",\"url\":\"https:\/\/www.neuraldesigner.com\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.neuraldesigner.com\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2023\/05\/logo-neural-1.png\",\"contentUrl\":\"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2023\/05\/logo-neural-1.png\",\"width\":1024,\"height\":223,\"caption\":\"Neural Designer\"},\"image\":{\"@id\":\"https:\/\/www.neuraldesigner.com\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/x.com\/NeuralDesigner\",\"https:\/\/es.linkedin.com\/showcase\/neuraldesigner\/\"]}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Melanoma classification from dermoscopic images - Neural Designer","description":"Classify 900 ISIC 2016 dermoscopic images as benign or melanoma with a convolutional network in Neural Designer, from data preparation to testing analysis.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/","og_locale":"en_US","og_type":"article","og_title":"Melanoma classification from dermoscopic images - Neural Designer","og_description":"Classify 900 ISIC 2016 dermoscopic images as benign or melanoma with a convolutional network in Neural Designer, from data preparation to testing analysis.","og_url":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/","og_site_name":"Neural Designer","article_modified_time":"2026-10-06T17:13:22+00:00","og_image":[{"url":"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg","type":"","width":"","height":""}],"twitter_card":"summary_large_image","twitter_site":"@NeuralDesigner","twitter_misc":{"Est. reading time":"9 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/","url":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/","name":"Melanoma classification from dermoscopic images - Neural Designer","isPartOf":{"@id":"https:\/\/www.neuraldesigner.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#primaryimage"},"image":{"@id":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#primaryimage"},"thumbnailUrl":"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg","datePublished":"2026-09-24T11:19:09+00:00","dateModified":"2026-10-06T17:13:22+00:00","description":"Classify 900 ISIC 2016 dermoscopic images as benign or melanoma with a convolutional network in Neural Designer, from data preparation to testing analysis.","breadcrumb":{"@id":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#primaryimage","url":"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg","contentUrl":"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2026\/10\/nd-melanoma900-samples-20261006.jpg","width":996,"height":492},{"@type":"BreadcrumbList","@id":"https:\/\/www.neuraldesigner.com\/learning\/examples\/melanoma-image-classification\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.neuraldesigner.com\/"},{"@type":"ListItem","position":2,"name":"Learning","item":"https:\/\/www.neuraldesigner.com\/learning\/"},{"@type":"ListItem","position":3,"name":"Melanoma classification from dermoscopic images"}]},{"@type":"WebSite","@id":"https:\/\/www.neuraldesigner.com\/#website","url":"https:\/\/www.neuraldesigner.com\/","name":"Neural Designer","description":"No-code AI Platform","publisher":{"@id":"https:\/\/www.neuraldesigner.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.neuraldesigner.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/www.neuraldesigner.com\/#organization","name":"Neural Designer","url":"https:\/\/www.neuraldesigner.com\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.neuraldesigner.com\/#\/schema\/logo\/image\/","url":"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2023\/05\/logo-neural-1.png","contentUrl":"https:\/\/www.neuraldesigner.com\/wp-content\/uploads\/2023\/05\/logo-neural-1.png","width":1024,"height":223,"caption":"Neural Designer"},"image":{"@id":"https:\/\/www.neuraldesigner.com\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/NeuralDesigner","https:\/\/es.linkedin.com\/showcase\/neuraldesigner\/"]}]}},"_links":{"self":[{"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/learning\/24127","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/learning"}],"about":[{"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/types\/learning"}],"author":[{"embeddable":true,"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/users\/152"}],"version-history":[{"count":5,"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/learning\/24127\/revisions"}],"predecessor-version":[{"id":24327,"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/learning\/24127\/revisions\/24327"}],"wp:attachment":[{"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/media?parent=24127"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/categories?post=24127"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.neuraldesigner.com\/api\/wp\/v2\/tags?post=24127"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}