{"id":1337,"date":"2021-01-19T13:44:32","date_gmt":"2021-01-19T15:44:32","guid":{"rendered":"https:\/\/sbia.org.br\/lnlm\/?page_id=1337"},"modified":"2021-01-19T13:44:32","modified_gmt":"2021-01-19T15:44:32","slug":"vol18-no2-art5","status":"publish","type":"page","link":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/","title":{"rendered":"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario"},"content":{"rendered":"<p>Marcelo Romero <a href=\"http:\/\/orcid.org\/0000-0002-8673-3744\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-1167\" src=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\" alt=\"orcid\" width=\"20\" height=\"20\" \/><\/a>, Matheus Gutoski <a href=\"http:\/\/orcid.org\/00000-0001-7679-0588\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-1167\" src=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\" alt=\"orcid\" width=\"20\" height=\"20\" \/><\/a>, Leandro Takeshi Hattori <a href=\"http:\/\/orcid.org\/0000-0002-1945-6855\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-1167\" src=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\" alt=\"orcid\" width=\"20\" height=\"20\" \/><\/a>, Manass\u00e9s Ribeiro <a href=\"http:\/\/orcid.org\/0000-0002-7526-5092\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-1167\" src=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\" alt=\"orcid\" width=\"20\" height=\"20\" \/><\/a> &amp; Heitor Silv\u00e9rio Lopes <a href=\"http:\/\/orcid.org\/0000-0003-3984-1432\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-1167\" src=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\" alt=\"orcid\" width=\"20\" height=\"20\" \/><\/a><\/p>\n<p><strong>Abstract:<\/strong> Transfer learning is a paradigm that consists in training and testing classifiers with datasets drawn from distinct distributions. This technique allows to solve a particular problem using a model that was trained for another purpose. In the recent years, this practice has become very popular due to the increase of public available pre-trained models that can be fine-tuned to be applied in different scenarios. However, the relationship between the datasets used for training the model and the test data is usually not addressed, specially where the fine-tuning process is done only for the fully connected layers of a Convolutional Neural Network with pre-trained weights. This work presents a study regarding the relationship between the datasets used in a transfer learning process in terms of the performance achieved by models complexities and similarities. For this purpose, we fine-tune the final layer of Convolutional Neural Networks with pre-trained weights using diverse soft biometrics datasets. An evaluation of the performances of the models when tested with datasets that are different from the one used for training the model is presented. Complexity and similarity metrics are also used to perform the evaluation.<\/p>\n<p><strong>Keywords:<\/strong> Neural Network, Convolutional Neural Network, Transfer Learning, Soft Biometrics, Data Complexity, Data Similarity.<\/p>\n<p><strong>DOI code:<\/strong> <a href=\"http:\/\/dx.doi.org\/10.21528\/lnlm-vol18-no2-art5\">10.21528\/lnlm-vol18-no2-art5<\/a><\/p>\n<p><strong>PDF file:<\/strong> <a href=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2021\/01\/vol18-no2-art5.pdf\">vol18-no2-art5.pdf<\/a><\/p>\n<p><strong>BibTex file:<\/strong> <a href=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2021\/01\/vol18-no2-art5.bib\">vol18-no2-art5.bib<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Marcelo Romero , Matheus Gutoski , Leandro Takeshi Hattori , Manass\u00e9s Ribeiro &amp; Heitor Silv\u00e9rio Lopes Abstract: Transfer learning is a paradigm that consists in training and testing classifiers with datasets drawn from distinct distributions. This technique allows to solve <a href=\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/\" class=\"read-more\">Read More &#8230;<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"parent":1306,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-1337","page","type-page","status-publish","hentry"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.9 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario - Learning and NonLinear Models<\/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:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/\" \/>\n<meta property=\"og:locale\" content=\"pt_BR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario - Learning and NonLinear Models\" \/>\n<meta property=\"og:description\" content=\"Marcelo Romero , Matheus Gutoski , Leandro Takeshi Hattori , Manass\u00e9s Ribeiro &amp; Heitor Silv\u00e9rio Lopes Abstract: Transfer learning is a paradigm that consists in training and testing classifiers with datasets drawn from distinct distributions. This technique allows to solve Read More ...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/\" \/>\n<meta property=\"og:site_name\" content=\"Learning and NonLinear Models\" \/>\n<meta property=\"og:image\" content=\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Est. tempo de leitura\" \/>\n\t<meta name=\"twitter:data1\" content=\"1 minuto\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/\",\"url\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/\",\"name\":\"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario - Learning and NonLinear Models\",\"isPartOf\":{\"@id\":\"https:\/\/sbia.org.br\/lnlm\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\",\"datePublished\":\"2021-01-19T15:44:32+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#breadcrumb\"},\"inLanguage\":\"pt-BR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"pt-BR\",\"@id\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#primaryimage\",\"url\":\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\",\"contentUrl\":\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Browse issues\",\"item\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Learning &#038; Nonlinear Models &#8211; L&#038;NLM &#8211; Volume 18 &#8211; N\u00famero 2\",\"item\":\"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/\"},{\"@type\":\"ListItem\",\"position\":3,\"name\":\"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/sbia.org.br\/lnlm\/#website\",\"url\":\"https:\/\/sbia.org.br\/lnlm\/\",\"name\":\"Learning and NonLinear Models\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\/\/sbia.org.br\/lnlm\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/sbia.org.br\/lnlm\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"pt-BR\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/sbia.org.br\/lnlm\/#organization\",\"name\":\"Learning and NonLinear Models\",\"url\":\"https:\/\/sbia.org.br\/lnlm\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"pt-BR\",\"@id\":\"https:\/\/sbia.org.br\/lnlm\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/2021\/07\/logo-lnlm.png\",\"contentUrl\":\"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/2021\/07\/logo-lnlm.png\",\"width\":398,\"height\":94,\"caption\":\"Learning and NonLinear Models\"},\"image\":{\"@id\":\"https:\/\/sbia.org.br\/lnlm\/#\/schema\/logo\/image\/\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario - Learning and NonLinear Models","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:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/","og_locale":"pt_BR","og_type":"article","og_title":"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario - Learning and NonLinear Models","og_description":"Marcelo Romero , Matheus Gutoski , Leandro Takeshi Hattori , Manass\u00e9s Ribeiro &amp; Heitor Silv\u00e9rio Lopes Abstract: Transfer learning is a paradigm that consists in training and testing classifiers with datasets drawn from distinct distributions. This technique allows to solve Read More ...","og_url":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/","og_site_name":"Learning and NonLinear Models","og_image":[{"url":"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg","type":"","width":"","height":""}],"twitter_card":"summary_large_image","twitter_misc":{"Est. tempo de leitura":"1 minuto"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/","url":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/","name":"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario - Learning and NonLinear Models","isPartOf":{"@id":"https:\/\/sbia.org.br\/lnlm\/#website"},"primaryImageOfPage":{"@id":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#primaryimage"},"image":{"@id":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#primaryimage"},"thumbnailUrl":"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg","datePublished":"2021-01-19T15:44:32+00:00","breadcrumb":{"@id":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#breadcrumb"},"inLanguage":"pt-BR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/"]}]},{"@type":"ImageObject","inLanguage":"pt-BR","@id":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#primaryimage","url":"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg","contentUrl":"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/sites\/4\/2020\/09\/orcid.jpg"},{"@type":"BreadcrumbList","@id":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/vol18-no2-art5\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Browse issues","item":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/"},{"@type":"ListItem","position":2,"name":"Learning &#038; Nonlinear Models &#8211; L&#038;NLM &#8211; Volume 18 &#8211; N\u00famero 2","item":"https:\/\/sbia.org.br\/lnlm\/publicacoes\/vol18-no2\/"},{"@type":"ListItem","position":3,"name":"A Study of the Influence of Data Complexity and Similarity on Soft Biometrics Classification Performance in a Transfer Learning Scenario"}]},{"@type":"WebSite","@id":"https:\/\/sbia.org.br\/lnlm\/#website","url":"https:\/\/sbia.org.br\/lnlm\/","name":"Learning and NonLinear Models","description":"","publisher":{"@id":"https:\/\/sbia.org.br\/lnlm\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/sbia.org.br\/lnlm\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"pt-BR"},{"@type":"Organization","@id":"https:\/\/sbia.org.br\/lnlm\/#organization","name":"Learning and NonLinear Models","url":"https:\/\/sbia.org.br\/lnlm\/","logo":{"@type":"ImageObject","inLanguage":"pt-BR","@id":"https:\/\/sbia.org.br\/lnlm\/#\/schema\/logo\/image\/","url":"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/2021\/07\/logo-lnlm.png","contentUrl":"https:\/\/sbia.org.br\/lnlm\/wp-content\/uploads\/2021\/07\/logo-lnlm.png","width":398,"height":94,"caption":"Learning and NonLinear Models"},"image":{"@id":"https:\/\/sbia.org.br\/lnlm\/#\/schema\/logo\/image\/"}}]}},"_links":{"self":[{"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/pages\/1337","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/comments?post=1337"}],"version-history":[{"count":0,"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/pages\/1337\/revisions"}],"up":[{"embeddable":true,"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/pages\/1306"}],"wp:attachment":[{"href":"https:\/\/sbia.org.br\/lnlm\/wp-json\/wp\/v2\/media?parent=1337"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}