{"id":866,"date":"2026-08-26T09:00:00","date_gmt":"2026-08-26T08:00:00","guid":{"rendered":"https:\/\/research.reading.ac.uk\/einstein\/?p=866"},"modified":"2026-08-26T09:00:00","modified_gmt":"2026-08-26T08:00:00","slug":"training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai","status":"publish","type":"post","link":"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/","title":{"rendered":"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI"},"content":{"rendered":"\n
\n \"Illustration\n<\/figure>\n\n\n\n

When a new face recognition algorithm claims state-of-the-art performance, or a morphing attack detector reports a near-zero error rate, the natural question is: state-of-the-art on what? Error rate on which dataset, evaluated under which conditions, by whom? These questions matter enormously. The performance of a biometric AI system is inseparable from the data it was trained on and the benchmark it was evaluated against.<\/p>\n\n\n\n

Why biometric datasets are hard to create<\/h2>\n\n\n\n

A useful dataset for training a morphing attack detector needs to contain both genuine (bona fide) facial images and morphed images \u2014 lots of them, across a wide range of morphing techniques, spanning diverse demographics, with realistic variation in image quality and capture conditions. Biometric data is special category personal data under GDPR, requiring explicit consent and careful data governance.<\/p>\n\n\n\n

The evaluation platform challenge<\/h2>\n\n\n\n

Beyond training data, the field needs independent evaluation platforms \u2014 systems that allow researchers to submit algorithms and receive performance metrics on held-out test data that the algorithm developers have never seen. The Bologna Online Evaluation Platform (BOEP) provides exactly this for morphing attack detection. NIST runs similar evaluation programmes for face recognition (FRVT), face image quality (SIDD), and morphing attack detection (FATE-MORPH).<\/p>\n\n\n\n

The ENCHANTER framework<\/h2>\n\n\n\n

The ENCHANTER framework, developed within the EINSTEIN project and released as open source, addresses the practical challenge of standardising dataset preparation pipelines \u2014 handling the heterogeneous formats, quality levels, and metadata conventions of different data sources. By reducing the engineering overhead of dataset preparation, ENCHANTER allows researchers to focus on algorithm development and makes it easier to reproduce and compare results across different groups.<\/p>\n\n\n\n

Open, independent evaluation is how the field measures genuine scientific progress. A biometric algorithm can only be as good as the data it was trained on and the benchmark it was evaluated against.<\/p>\n\n\n\n

© 2026 EINSTEIN Consortium. EINSTEIN is funded by the European Union’s Horizon Europe programme (GA No. 101121280) and by UKRI (IFS 10093453). Views expressed are those of the authors only. www.einstein-horizon.eu<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

When a new face recognition algorithm claims state-of-the-art performance, or a morphing attack detector reports a near-zero error rate, the natural question is: state-of-the-art on what? Error rate on which dataset, evaluated under which conditions,…<\/p>\n","protected":false},"author":909,"featured_media":832,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"__cvm_playback_settings":[],"__cvm_video_id":"","footnotes":""},"categories":[18],"tags":[90,19,92,89,94,21,93,91,24],"coauthors":[14],"class_list":["post-866","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-research-blog","tag-benchmarks","tag-biometrics","tag-boep","tag-datasets","tag-evaluation","tag-face-recognition","tag-machine-learning","tag-nist","tag-pad"],"acf":[],"yoast_head":"\nTraining the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI - EINSTEIN<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/\" \/>\n<meta property=\"og:locale\" content=\"en_GB\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI - EINSTEIN\" \/>\n<meta property=\"og:description\" content=\"When a new face recognition algorithm claims state-of-the-art performance, or a morphing attack detector reports a near-zero error rate, the natural question is: state-of-the-art on what? Error rate on which dataset, evaluated under which conditions,...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/\" \/>\n<meta property=\"og:site_name\" content=\"EINSTEIN\" \/>\n<meta property=\"article:published_time\" content=\"2026-08-26T08:00:00+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/research.reading.ac.uk\/einstein\/wp-content\/uploads\/sites\/327\/2026\/05\/einstein_blog_banner-1.png\" \/>\n\t<meta property=\"og:image:width\" content=\"720\" \/>\n\t<meta property=\"og:image:height\" content=\"405\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"James Ferryman\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"James Ferryman\" \/>\n\t<meta name=\"twitter:label2\" content=\"Estimated reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"2 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/\",\"url\":\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/\",\"name\":\"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI - EINSTEIN\",\"isPartOf\":{\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/#website\"},\"datePublished\":\"2026-08-26T08:00:00+00:00\",\"dateModified\":\"2026-08-26T08:00:00+00:00\",\"author\":{\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/#\/schema\/person\/287e525ca50fc2213a8cd2ad36fb324f\"},\"breadcrumb\":{\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/#breadcrumb\"},\"inLanguage\":\"en-GB\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/research.reading.ac.uk\/einstein\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/#website\",\"url\":\"https:\/\/research.reading.ac.uk\/einstein\/\",\"name\":\"EINSTEIN\",\"description\":\"IntEroperable applIcations suIte to eNhance European identity and document SecuriTy and fraud dEtectIoN\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/research.reading.ac.uk\/einstein\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-GB\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/#\/schema\/person\/287e525ca50fc2213a8cd2ad36fb324f\",\"name\":\"James Ferryman\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-GB\",\"@id\":\"https:\/\/research.reading.ac.uk\/einstein\/#\/schema\/person\/image\/e455eff97bf15f08533b5794a7ae2be3\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/bb9b6c1740f6a54e1761096e29ce475824a3bb90f2b70ebc385aa96656998cc9?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/bb9b6c1740f6a54e1761096e29ce475824a3bb90f2b70ebc385aa96656998cc9?s=96&d=mm&r=g\",\"caption\":\"James Ferryman\"},\"url\":\"https:\/\/research.reading.ac.uk\/einstein\/author\/sis00jf\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI - EINSTEIN","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:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/","og_locale":"en_GB","og_type":"article","og_title":"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI - EINSTEIN","og_description":"When a new face recognition algorithm claims state-of-the-art performance, or a morphing attack detector reports a near-zero error rate, the natural question is: state-of-the-art on what? Error rate on which dataset, evaluated under which conditions,...","og_url":"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/","og_site_name":"EINSTEIN","article_published_time":"2026-08-26T08:00:00+00:00","og_image":[{"width":720,"height":405,"url":"https:\/\/research.reading.ac.uk\/einstein\/wp-content\/uploads\/sites\/327\/2026\/05\/einstein_blog_banner-1.png","type":"image\/png"}],"author":"James Ferryman","twitter_card":"summary_large_image","twitter_misc":{"Written by":"James Ferryman","Estimated reading time":"2 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/","url":"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/","name":"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI - EINSTEIN","isPartOf":{"@id":"https:\/\/research.reading.ac.uk\/einstein\/#website"},"datePublished":"2026-08-26T08:00:00+00:00","dateModified":"2026-08-26T08:00:00+00:00","author":{"@id":"https:\/\/research.reading.ac.uk\/einstein\/#\/schema\/person\/287e525ca50fc2213a8cd2ad36fb324f"},"breadcrumb":{"@id":"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/#breadcrumb"},"inLanguage":"en-GB","potentialAction":[{"@type":"ReadAction","target":["https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/research.reading.ac.uk\/einstein\/training-the-machine-why-datasets-and-benchmarks-are-the-unsung-heroes-of-biometric-ai\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/research.reading.ac.uk\/einstein\/"},{"@type":"ListItem","position":2,"name":"Training the Machine: Why Datasets and Benchmarks Are the Unsung Heroes of Biometric AI"}]},{"@type":"WebSite","@id":"https:\/\/research.reading.ac.uk\/einstein\/#website","url":"https:\/\/research.reading.ac.uk\/einstein\/","name":"EINSTEIN","description":"IntEroperable applIcations suIte to eNhance European identity and document SecuriTy and fraud dEtectIoN","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/research.reading.ac.uk\/einstein\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-GB"},{"@type":"Person","@id":"https:\/\/research.reading.ac.uk\/einstein\/#\/schema\/person\/287e525ca50fc2213a8cd2ad36fb324f","name":"James Ferryman","image":{"@type":"ImageObject","inLanguage":"en-GB","@id":"https:\/\/research.reading.ac.uk\/einstein\/#\/schema\/person\/image\/e455eff97bf15f08533b5794a7ae2be3","url":"https:\/\/secure.gravatar.com\/avatar\/bb9b6c1740f6a54e1761096e29ce475824a3bb90f2b70ebc385aa96656998cc9?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/bb9b6c1740f6a54e1761096e29ce475824a3bb90f2b70ebc385aa96656998cc9?s=96&d=mm&r=g","caption":"James Ferryman"},"url":"https:\/\/research.reading.ac.uk\/einstein\/author\/sis00jf\/"}]}},"_links":{"self":[{"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/posts\/866","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/users\/909"}],"replies":[{"embeddable":true,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/comments?post=866"}],"version-history":[{"count":1,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/posts\/866\/revisions"}],"predecessor-version":[{"id":1000,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/posts\/866\/revisions\/1000"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/media\/832"}],"wp:attachment":[{"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/media?parent=866"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/categories?post=866"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/tags?post=866"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/research.reading.ac.uk\/einstein\/wp-json\/wp\/v2\/coauthors?post=866"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}