{"id":182687,"date":"2008-02-12T00:00:00","date_gmt":"2009-10-31T09:54:26","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/msr-research-item\/candidate-talk-a-discriminative-kernel-based-model-to-rank-images-from-text-queries\/"},"modified":"2016-09-09T09:52:26","modified_gmt":"2016-09-09T16:52:26","slug":"candidate-talk-a-discriminative-kernel-based-model-to-rank-images-from-text-queries","status":"publish","type":"msr-video","link":"https:\/\/www.microsoft.com\/en-us\/research\/video\/candidate-talk-a-discriminative-kernel-based-model-to-rank-images-from-text-queries\/","title":{"rendered":"Candidate Talk: A Discriminative Kernel-based Model to Rank Images from Text Queries"},"content":{"rendered":"<div class=\"asset-content\">\n<p>This  presentation introduces a discriminative model for the retrieval of pictures from text queries. The core idea of this approach is to minimize a loss directly related to the retrieval performance of the model. For that purpose, we formalize the retrieval task as a ranking problem, and introduce a learning procedure optimizing a loss function related to the ranking performance. This strategy hence addresses the retrieval problem directly and does not rely on an intermediate image annotation task, which contrasts with previous research. Moreover, our learning procedure builds upon recent work on the online learning of kernel- based classifiers. This yields an efficient, scalable algorithm, which can benefit from recent kernels developed for image comparison.<br \/>\nThe experiments performed over stock photography data show the advantage of our discriminative ranking approach over state-of-the- art alternatives (e.g. our model yields 26.3% average precision over the Corel dataset, which should be compared to 22.0%, for the best alternative model evaluated). Further analysis of the results shows that our model is especially advantageous over difficult queries such as queries with few relevant pictures or multiple-word queries.<\/p>\n<\/div>\n<p><!-- .asset-content --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>This presentation introduces a discriminative model for the retrieval of pictures from text queries. The core idea of this approach is to minimize a loss directly related to the retrieval performance of the model. For that purpose, we formalize the retrieval task as a ranking problem, and introduce a learning procedure optimizing a loss function [&hellip;]<\/p>\n","protected":false},"featured_media":194733,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr_hide_image_in_river":0,"footnotes":""},"research-area":[],"msr-video-type":[],"msr-locale":[268875],"msr-post-option":[],"msr-session-type":[],"msr-impact-theme":[],"msr-pillar":[],"msr-episode":[],"msr-research-theme":[],"class_list":["post-182687","msr-video","type-msr-video","status-publish","has-post-thumbnail","hentry","msr-locale-en_us"],"msr_download_urls":"","msr_external_url":"https:\/\/youtu.be\/m4VY6QWAwHM","msr_secondary_video_url":"","msr_video_file":"","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video\/182687","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-video"}],"version-history":[{"count":0,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video\/182687\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/194733"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=182687"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=182687"},{"taxonomy":"msr-video-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-video-type?post=182687"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=182687"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=182687"},{"taxonomy":"msr-session-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-session-type?post=182687"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=182687"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=182687"},{"taxonomy":"msr-episode","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-episode?post=182687"},{"taxonomy":"msr-research-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-theme?post=182687"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}