{"id":350828,"date":"2017-01-11T11:28:09","date_gmt":"2017-01-11T19:28:09","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/?post_type=msr-research-item&p=350828"},"modified":"2018-10-16T19:55:50","modified_gmt":"2018-10-17T02:55:50","slug":"salient-object-detection-discriminative-regional-featureintegration-approach","status":"publish","type":"msr-research-item","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/salient-object-detection-discriminative-regional-featureintegration-approach\/","title":{"rendered":"Salient Object Detection: A Discriminative Regional FeatureIntegration Approach"},"content":{"rendered":"
Feature<\/span> inte<\/span>gration<\/span> provides<\/span> a<\/span> computational<\/span><\/span><\/div>\n
frame<\/span>work<\/span> for<\/span> saliency<\/span> detection,<\/span> and<\/span> a<\/span> lot<\/span> of<\/span> hand-crafted<\/span><\/div>\n
inte<\/span>gration<\/span> rules<\/span> ha<\/span>ve<\/span> been<\/span> de<\/span>veloped.<\/span> In<\/span> this<\/span> paper<\/span>,<\/span> we<\/span><\/div>\n
present<\/span> a<\/span> principled<\/span> e<\/span>xtension,<\/span> supervised<\/span> feature<\/span> integra-<\/span><\/div>\n
tion,<\/span> which<\/span> learns<\/span> a<\/span> random<\/span> forest<\/span> re<\/span>gressor<\/span> to<\/span> discrimina-<\/span><\/div>\n
ti<\/span>v<\/span>ely<\/span> integrate<\/span> the<\/span> salienc<\/span>y<\/span> features<\/span> for<\/span> salienc<\/span>y<\/span> computa-<\/span><\/div>\n
tion.<\/span> In<\/span> addition<\/span> to<\/span> contrast<\/span> features,<\/span> we<\/span> introduce<\/span> re<\/span>gional<\/span><\/div>\n
object-sensiti<\/span>v<\/span>e<\/span> descriptors:<\/span> the<\/span> objectness<\/span> descriptor<\/span> char<\/span>–<\/span><\/div>\n
acterizing<\/span> the<\/span> common<\/span> spatial<\/span> and<\/span> appearance<\/span> property<\/span> of<\/span><\/div>\n
the<\/span> salient<\/span> object,<\/span> and<\/span> the<\/span> image-speci<\/span>\ufb01<\/span>c<\/span> backgroundness<\/span><\/div>\n
descriptor<\/span> characterizing<\/span> the<\/span> appearance<\/span> of<\/span> the<\/span> background<\/span><\/div>\n
of<\/span> a<\/span> speci<\/span>\ufb01<\/span>c<\/span> image,<\/span> which<\/span> are<\/span> sho<\/span>wn<\/span> more<\/span> important<\/span> for<\/span><\/div>\n
estimating<\/span> the<\/span> salienc<\/span>y<\/span>.<\/span> T<\/span>o<\/span> the<\/span> best<\/span> of<\/span> our<\/span> kno<\/span>wledge,<\/span> our<\/span><\/div>\n
supervised<\/span> feature<\/span> integration<\/span> frame<\/span>work<\/span> is<\/span> the<\/span> \ufb01<\/span>rst<\/span> success-<\/span><\/div>\n
ful<\/span> approach<\/span> to<\/span> perform<\/span> the<\/span> integration<\/span> ov<\/span>er<\/span> the<\/span> saliency<\/span><\/div>\n
features<\/span> for<\/span> salient<\/span> object<\/span> detection,<\/span> and<\/span> outperforms<\/span> the<\/span><\/div>\n
inte<\/span>gration<\/span> approach<\/span> o<\/span>ver<\/span> the<\/span> salienc<\/span>y<\/span> maps.<\/span> T<\/span>ogether<\/span> with<\/span><\/div>\n
fusing<\/span> the<\/span> multi-le<\/span>v<\/span>el<\/span> regional<\/span> salienc<\/span>y<\/span> maps<\/span> to<\/span> impose<\/span><\/div>\n
the<\/span> spatial<\/span> salienc<\/span>y<\/span> consistenc<\/span>y<\/span>,<\/span> our<\/span> approach<\/span> signi<\/span>\ufb01<\/span>cantly<\/span><\/div>\n
outperforms<\/span> state-of-the-art<\/span> methods<\/span> on<\/span> se<\/span>ven<\/span> benchmark<\/span><\/div>\n
datasets.<\/span> W<\/span>e<\/span> also<\/span> discuss<\/span> se<\/span>veral<\/span> follo<\/span>wup<\/span> works<\/span> which\u00a0<\/span><\/span><\/p>\n
jointly<\/span> learn<\/span> the<\/span> representation<\/span> and<\/span> the<\/span> salienc<\/span>y<\/span> map<\/span> using<\/span><\/div>\n
deep<\/span> learning.<\/span><\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

Feature integration provides a computational framework for saliency detection, and a lot of hand-crafted integration rules have been developed. In this paper, we present a principled extension, supervised feature integra- tion, which learns a random forest regressor to discrimina- tively integrate the saliency features for saliency computa- tion. In addition to contrast features, we introduce […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":null,"msr_publishername":"Springer","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"","msr_doi":"","msr_arxiv_id":"","msr_s2_paper_id":"","msr_mag_id":"","msr_pubmed_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_original_fields_of_study":"","msr_release_tracker_id":"","msr_s2_match_type":"","msr_citation_count_updated":"","msr_published_date":"2016-11-19","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"http:\/\/www.readcube.com\/articles\/10.1007\/s11263-016-0977-3?author_access_token=kOMMcjAcMf08StVk9j_ChPe4RwlQNchNByi7wbcMAY6bGK4nxWGm7p1_IwKxtD7cM55EhD3uI8Yr3Bf64FyCfVfipG0la-Rwo5Oh-ia9qwwutitz7MV5IkN1PU39f0-_eBWAgnk9bZrtzosTBnAX-A%3D%3D","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_s2_pdf_url":"","msr_year":0,"msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_match_confidence":0,"msr_microsoftintellectualproperty":true,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13563],"msr-publication-type":[193715],"msr-publisher":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-350828","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-data-platform-analytics","msr-locale-en_us"],"msr_publishername":"Springer","msr_edition":"","msr_affiliation":"","msr_published_date":"2016-11-19","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"","msr_publicationurl":"http:\/\/www.readcube.com\/articles\/10.1007\/s11263-016-0977-3?author_access_token=kOMMcjAcMf08StVk9j_ChPe4RwlQNchNByi7wbcMAY6bGK4nxWGm7p1_IwKxtD7cM55EhD3uI8Yr3Bf64FyCfVfipG0la-Rwo5Oh-ia9qwwutitz7MV5IkN1PU39f0-_eBWAgnk9bZrtzosTBnAX-A%3D%3D","msr_doi":"","msr_publication_uploader":[{"type":"url","title":"http:\/\/www.readcube.com\/articles\/10.1007\/s11263-016-0977-3?author_access_token=kOMMcjAcMf08StVk9j_ChPe4RwlQNchNByi7wbcMAY6bGK4nxWGm7p1_IwKxtD7cM55EhD3uI8Yr3Bf64FyCfVfipG0la-Rwo5Oh-ia9qwwutitz7MV5IkN1PU39f0-_eBWAgnk9bZrtzosTBnAX-A%3D%3D","viewUrl":false,"id":false,"label_id":0}],"msr_related_uploader":"","msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[{"id":0,"url":"http:\/\/www.readcube.com\/articles\/10.1007\/s11263-016-0977-3?author_access_token=kOMMcjAcMf08StVk9j_ChPe4RwlQNchNByi7wbcMAY6bGK4nxWGm7p1_IwKxtD7cM55EhD3uI8Yr3Bf64FyCfVfipG0la-Rwo5Oh-ia9qwwutitz7MV5IkN1PU39f0-_eBWAgnk9bZrtzosTBnAX-A%3D%3D"}],"msr-author-ordering":[{"type":"user_nicename","value":"jingdw","user_id":32299,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=jingdw"},{"type":"text","value":"Huaizu Jiang","user_id":0,"rest_url":false},{"type":"text","value":"Zejian Yuan","user_id":0,"rest_url":false},{"type":"text","value":"Ming-Ming Cheng","user_id":0,"rest_url":false},{"type":"text","value":"Xiaowei Hu","user_id":0,"rest_url":false},{"type":"text","value":"Nanning Zheng","user_id":0,"rest_url":false}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[],"msr_project":[],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"article","related_content":[],"_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/350828","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":1,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/350828\/revisions"}],"predecessor-version":[{"id":512693,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/350828\/revisions\/512693"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=350828"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=350828"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=350828"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=350828"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=350828"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=350828"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=350828"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=350828"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=350828"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=350828"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=350828"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=350828"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=350828"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}