{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/16772"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/16772","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Joint appearance and locality image representation by Gaussianization","abstract":"A novel image representation is proposed in this thesis to capture both the appearance and locality information for image classification applications. First, we model the feature vectors, from various granularity levels including the corpus level, the image level and image patch level, in a hierarchical Bayesian framework using mixtures of Gaussians. After such a hierarchical Gaussianization, each image is represented as a Gaussian mixture model (GMM) for its appearance, and several Gaussian maps for its spatial layout. Then we extract the appearance information from the GMM parameters, and the locality information from the global and the local statistics over Gaussian maps. Finally, we employ a supervised dimension reduction technique called DAP (discriminant adaptive projection) to remove noise directions and to further enhance the discriminating power of our representation. To validate the argument that the new representation is a general representation for images and video frames, we evaluate the representation on several important applications. Firstly, we apply the new presentation to classification and regression tasks taking whole images as inputs. These tasks include object recognition, scene category classification, face recognition, age estimation, pose estimation, gender recognition, and video event recognition. Then we test it for the object detection and image parsing tasks, where the new representation takes partial images as inputs. The experimental results show that, for various types of images and tasks, the performances using the proposed representation were the best in all the applications compared with other state-of-the-art algorithms.","abstract_html":"A novel image representation is proposed in this thesis to capture both the appearance and locality information for image classification applications. First, we model the feature vectors, from various granularity levels including the corpus level, the image level and image patch level, in a hierarchical Bayesian framework using mixtures of Gaussians. After such a hierarchical Gaussianization, each image is represented as a Gaussian mixture model (GMM) for its appearance, and several Gaussian maps for its spatial layout. Then we extract the appearance information from the GMM parameters, and the locality information from the global and the local statistics over Gaussian maps. Finally, we employ a supervised dimension reduction technique called DAP (discriminant adaptive projection) to remove noise directions and to further enhance the discriminating power of our representation. To validate the argument that the new representation is a general representation for images and video frames, we evaluate the representation on several important applications. Firstly, we apply the new presentation to classification and regression tasks taking whole images as inputs. These tasks include object recognition, scene category classification, face recognition, age estimation, pose estimation, gender recognition, and video event recognition. Then we test it for the object detection and image parsing tasks, where the new representation takes partial images as inputs. The experimental results show that, for various types of images and tasks, the performances using the proposed representation were the best in all the applications compared with other state-of-the-art algorithms.","abstract_has_math":false,"creators":["Zhou, Xi"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Electrical & Computer Engr","degree_department":null,"school":null,"contributors":["Huang, Thomas S.","Hasegawa-Johnson, Mark A.","Levinson, Stephen E.","Liang, Feng"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2010,"date_issued":"2010-08-20T17:57:22Z","date_published":"2010-08-20T17:57:22Z","updated_at":"2026-07-22T22:25:09Z","subjects":["image representation","hierarchical Gaussianization","discriminant attribute projection"],"languages":["en"],"rights":["Copyright 2010 Xi Zhou"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/16772","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Huang, Thomas S.","Hasegawa-Johnson, Mark A.","Levinson, Stephen E.","Liang, Feng"]},{"key":"dc:creator","label":"Author","values":["Zhou, Xi"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2010-08-20T17:57:22Z","2010-08"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical & Computer Engr"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["image representation","hierarchical Gaussianization","discriminant attribute projection"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2010 Xi Zhou"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/16772"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["A novel image representation is proposed in this thesis to capture both the appearance and locality information for image classification applications. First, we model the feature vectors, from various granularity levels including the corpus level, the image level and image patch level, in a hierarchical Bayesian framework using mixtures of Gaussians. After such a hierarchical Gaussianization, each image is represented as a Gaussian mixture model (GMM) for its appearance, and several Gaussian maps for its spatial layout. Then we extract the appearance information from the GMM parameters, and the locality information from the global and the local statistics over Gaussian maps. Finally, we employ a supervised dimension reduction technique called DAP (discriminant adaptive projection) to remove noise directions and to further enhance the discriminating power of our representation. To validate the argument that the new representation is a general representation for images and video frames, we evaluate the representation on several important applications. Firstly, we apply the new presentation to classification and regression tasks taking whole images as inputs. These tasks include object recognition, scene category classification, face recognition, age estimation, pose estimation, gender recognition, and video event recognition. Then we test it for the object detection and image parsing tasks, where the new representation takes partial images as inputs. The experimental results show that, for various types of images and tasks, the performances using the proposed representation were the best in all the applications compared with other state-of-the-art algorithms.","Item withdrawn by Alexis Thompson (athmpsn1@illinois.edu) on 2010-07-12T21:36:23Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 Zhou_Xi.pdf: 8464059 bytes, checksum: 41f30cb0d7f031bb0e1850b8a1648d0d (MD5)","Made available in DSpace on 2010-08-20T17:57:22Z (GMT). No. of bitstreams: 3 Zhou_Xi.pdf: 8464059 bytes, checksum: 41f30cb0d7f031bb0e1850b8a1648d0d (MD5) 1_Zhou_Xi.pdf: 8464055 bytes, checksum: 89f0cbefac56b2952002b30f4b1e21a9 (MD5) license.txt: 4056 bytes, checksum: ae724011e4cbda6745bb5d61340e9318 (MD5)"]},{"key":"dc:title","label":"Title","values":["Joint appearance and locality image representation by Gaussianization"]}]}],"canonical_facts":{"dc:contributor":["Huang, Thomas S.","Hasegawa-Johnson, Mark A.","Levinson, Stephen E.","Liang, Feng"],"dc:creator":["Zhou, Xi"],"dc:date":["2010-08-20T17:57:22Z","2010-08"],"dc:description":["A novel image representation is proposed in this thesis to capture both the appearance and locality information for image classification applications. First, we model the feature vectors, from various granularity levels including the corpus level, the image level and image patch level, in a hierarchical Bayesian framework using mixtures of Gaussians. After such a hierarchical Gaussianization, each image is represented as a Gaussian mixture model (GMM) for its appearance, and several Gaussian maps for its spatial layout. Then we extract the appearance information from the GMM parameters, and the locality information from the global and the local statistics over Gaussian maps. Finally, we employ a supervised dimension reduction technique called DAP (discriminant adaptive projection) to remove noise directions and to further enhance the discriminating power of our representation. To validate the argument that the new representation is a general representation for images and video frames, we evaluate the representation on several important applications. Firstly, we apply the new presentation to classification and regression tasks taking whole images as inputs. These tasks include object recognition, scene category classification, face recognition, age estimation, pose estimation, gender recognition, and video event recognition. Then we test it for the object detection and image parsing tasks, where the new representation takes partial images as inputs. The experimental results show that, for various types of images and tasks, the performances using the proposed representation were the best in all the applications compared with other state-of-the-art algorithms.","Item withdrawn by Alexis Thompson (athmpsn1@illinois.edu) on 2010-07-12T21:36:23Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 Zhou_Xi.pdf: 8464059 bytes, checksum: 41f30cb0d7f031bb0e1850b8a1648d0d (MD5)","Made available in DSpace on 2010-08-20T17:57:22Z (GMT). No. of bitstreams: 3 Zhou_Xi.pdf: 8464059 bytes, checksum: 41f30cb0d7f031bb0e1850b8a1648d0d (MD5) 1_Zhou_Xi.pdf: 8464055 bytes, checksum: 89f0cbefac56b2952002b30f4b1e21a9 (MD5) license.txt: 4056 bytes, checksum: ae724011e4cbda6745bb5d61340e9318 (MD5)"],"dc:identifier":["http://hdl.handle.net/2142/16772"],"dc:language":["en"],"dc:rights":["Copyright 2010 Xi Zhou"],"dc:subject":["image representation","hierarchical Gaussianization","discriminant attribute projection"],"dc:title":["Joint appearance and locality image representation by Gaussianization"],"thesis:degree_discipline":["Electrical & Computer Engr"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:09Z"}