{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/18612"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/18612","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Hierarchical density estimation for image classification","abstract":"\"Histogram (bag-of-words) and Gaussian mixture models (GMMs) have been widely used in patch-based image classification problems. Despite the satisfactory results reported, both methods suffer from a number of disadvantages. For instance, a histogram may be easy to learn but has a large quantization error; on the contrary, Gaussian mixture model based methods have better modeling capabilities but are inefficient in both learning and testing. In this thesis, we present a novel hierarchical density estimation approach for image classification. This new approach partitions the feature space into small regions using a tree structure. For each region, \"\"local\"\" distribution is characterized by class-conditional Gaussians via hierarchical maximum a posteriori (MAP) estimation. We further enhance the parameter estimation by smoothing over a collection of randomized trees. This new approach enjoys the merits of superior modeling capability, robust parameter estimation, and efficient testing. Experiments on scene classification demonstrate both the effectiveness and efficiency of this new approach.\"","abstract_html":""Histogram (bag-of-words) and Gaussian mixture models (GMMs) have been widely used in patch-based image classification problems. Despite the satisfactory results reported, both methods suffer from a number of disadvantages. For instance, a histogram may be easy to learn but has a large quantization error; on the contrary, Gaussian mixture model based methods have better modeling capabilities but are inefficient in both learning and testing. In this thesis, we present a novel hierarchical density estimation approach for image classification. This new approach partitions the feature space into small regions using a tree structure. For each region, ""local"" distribution is characterized by class-conditional Gaussians via hierarchical maximum a posteriori (MAP) estimation. We further enhance the parameter estimation by smoothing over a collection of randomized trees. This new approach enjoys the merits of superior modeling capability, robust parameter estimation, and efficient testing. Experiments on scene classification demonstrate both the effectiveness and efficiency of this new approach."","abstract_has_math":false,"creators":["Li, Zhen"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Electrical & Computer Engr","degree_department":null,"school":null,"contributors":["Huang, Thomas S."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2011,"date_issued":"2011-01-21T22:51:40Z","date_published":"2011-01-21T22:51:40Z","updated_at":"2026-07-22T22:25:11Z","subjects":["Hierarchical maximum a posteriori (MAP) estimation","random forest","localized Gaussian"],"languages":["en"],"rights":["Copyright 2010 Zhen Li"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/18612","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Huang, Thomas S."]},{"key":"dc:creator","label":"Author","values":["Li, Zhen"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2011-01-21T22:51:40Z","2013-01-22T11:00:15Z","2010-12"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Electrical & Computer Engr"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M.S."]},{"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":["Hierarchical maximum a posteriori (MAP) estimation","random forest","localized Gaussian"]}]},{"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 Zhen Li"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/18612"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["\"Histogram (bag-of-words) and Gaussian mixture models (GMMs) have been widely used in patch-based image classification problems. Despite the satisfactory results reported, both methods suffer from a number of disadvantages. For instance, a histogram may be easy to learn but has a large quantization error; on the contrary, Gaussian mixture model based methods have better modeling capabilities but are inefficient in both learning and testing. In this thesis, we present a novel hierarchical density estimation approach for image classification. This new approach partitions the feature space into small regions using a tree structure. For each region, \"\"local\"\" distribution is characterized by class-conditional Gaussians via hierarchical maximum a posteriori (MAP) estimation. We further enhance the parameter estimation by smoothing over a collection of randomized trees. This new approach enjoys the merits of superior modeling capability, robust parameter estimation, and efficient testing. Experiments on scene classification demonstrate both the effectiveness and efficiency of this new approach.\"","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2010-11-29T21:41:20Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 LI_Zhen.pdf: 2574155 bytes, checksum: 5c7d556ad32043034ca2c8ad2adfdb4e (MD5)","Made available in DSpace on 2011-01-21T22:51:40Z (GMT). No. of bitstreams: 2 LI_Zhen.pdf: 2574155 bytes, checksum: 5c7d556ad32043034ca2c8ad2adfdb4e (MD5) license.txt: 4056 bytes, checksum: 7e4b73087fe60008aaed6f813e083ce5 (MD5)","Item marked as restricted to the 'Administrator' Group (id=1) by William Ingram (wingram2@illinois.edu) on 2011-01-21T22:53:51Z Item is restricted until 2013-01-21T22:53:34Z","Item reinstated by Sarah Shreeves (sshreeve@illinois.edu) on 2013-01-22T11:00:15Z Item was in collections: University of Illinois Dissertations and Theses (ID: 204) Dissertations and Theses - Electrical and Computer Engineering (ID: 446) No. of bitstreams: 3 LI_Zhen.pdf.txt: 33940 bytes, checksum: 510114d3e08c41c9de38fe64e1f6b3d8 (MD5) LI_Zhen.pdf: 2574155 bytes, checksum: 5c7d556ad32043034ca2c8ad2adfdb4e (MD5) license.txt: 4056 bytes, checksum: 7e4b73087fe60008aaed6f813e083ce5 (MD5)","Item released from any restrictions by Sarah Shreeves (sshreeve@illinois.edu) on 2013-01-22T11:00:15Z"]},{"key":"dc:title","label":"Title","values":["Hierarchical density estimation for image classification"]}]}],"canonical_facts":{"dc:contributor":["Huang, Thomas S."],"dc:creator":["Li, Zhen"],"dc:date":["2011-01-21T22:51:40Z","2013-01-22T11:00:15Z","2010-12"],"dc:description":["\"Histogram (bag-of-words) and Gaussian mixture models (GMMs) have been widely used in patch-based image classification problems. Despite the satisfactory results reported, both methods suffer from a number of disadvantages. For instance, a histogram may be easy to learn but has a large quantization error; on the contrary, Gaussian mixture model based methods have better modeling capabilities but are inefficient in both learning and testing. In this thesis, we present a novel hierarchical density estimation approach for image classification. This new approach partitions the feature space into small regions using a tree structure. For each region, \"\"local\"\" distribution is characterized by class-conditional Gaussians via hierarchical maximum a posteriori (MAP) estimation. We further enhance the parameter estimation by smoothing over a collection of randomized trees. This new approach enjoys the merits of superior modeling capability, robust parameter estimation, and efficient testing. Experiments on scene classification demonstrate both the effectiveness and efficiency of this new approach.\"","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2010-11-29T21:41:20Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 LI_Zhen.pdf: 2574155 bytes, checksum: 5c7d556ad32043034ca2c8ad2adfdb4e (MD5)","Made available in DSpace on 2011-01-21T22:51:40Z (GMT). No. of bitstreams: 2 LI_Zhen.pdf: 2574155 bytes, checksum: 5c7d556ad32043034ca2c8ad2adfdb4e (MD5) license.txt: 4056 bytes, checksum: 7e4b73087fe60008aaed6f813e083ce5 (MD5)","Item marked as restricted to the 'Administrator' Group (id=1) by William Ingram (wingram2@illinois.edu) on 2011-01-21T22:53:51Z Item is restricted until 2013-01-21T22:53:34Z","Item reinstated by Sarah Shreeves (sshreeve@illinois.edu) on 2013-01-22T11:00:15Z Item was in collections: University of Illinois Dissertations and Theses (ID: 204) Dissertations and Theses - Electrical and Computer Engineering (ID: 446) No. of bitstreams: 3 LI_Zhen.pdf.txt: 33940 bytes, checksum: 510114d3e08c41c9de38fe64e1f6b3d8 (MD5) LI_Zhen.pdf: 2574155 bytes, checksum: 5c7d556ad32043034ca2c8ad2adfdb4e (MD5) license.txt: 4056 bytes, checksum: 7e4b73087fe60008aaed6f813e083ce5 (MD5)","Item released from any restrictions by Sarah Shreeves (sshreeve@illinois.edu) on 2013-01-22T11:00:15Z"],"dc:identifier":["http://hdl.handle.net/2142/18612"],"dc:language":["en"],"dc:rights":["Copyright 2010 Zhen Li"],"dc:subject":["Hierarchical maximum a posteriori (MAP) estimation","random forest","localized Gaussian"],"dc:title":["Hierarchical density estimation for image classification"],"thesis:degree_discipline":["Electrical & Computer Engr"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:11Z"}