{"id":{"repo_id":"utc","oai_identifier":"oai:scholar.utc.edu:theses-1869"},"canonical_url":"https://search.dev.ndltd.org/etd/utc/oai:scholar.utc.edu:theses-1869","repository":{"repo_id":"utc","name":"University of Tennessee - Chattanooga","base_url":"https://scholar.utc.edu/do/oai/"},"display":{"title":"Deep learning-based fine-tuned multi-vehicle tracking with classification correction","abstract":"Single-camera multi-vehicle tracking is a preliminary step for traffic optimization. To perform vehicle tracking, a fine-tuned classification dataset was created from multiple sources. A ResNet classifier was trained, and its knowledge was utilized to build a localization dataset automatically without manual annotations. The dataset consists of videos generated from the MLK testbed. Three different object detection frameworks (SSD, Yolov3, and Yolov4) were evaluated with a pipeline that consists of the detector, the fine-tuned classifier and the tracker (DeepSort). Then the detector with the highest accuracy (Yolov3) was trained on the localization dataset. 90% of the ResNet classifier knowledge was distilled successfully in the final trained Yolov3 model. The tracker classification accuracy was further improved by proposing a correction methodology that considers both the camera's distance and limited data from the future frames.","abstract_html":"Single-camera multi-vehicle tracking is a preliminary step for traffic optimization. To perform vehicle tracking, a fine-tuned classification dataset was created from multiple sources. A ResNet classifier was trained, and its knowledge was utilized to build a localization dataset automatically without manual annotations. The dataset consists of videos generated from the MLK testbed. Three different object detection frameworks (SSD, Yolov3, and Yolov4) were evaluated with a pipeline that consists of the detector, the fine-tuned classifier and the tracker (DeepSort). Then the detector with the highest accuracy (Yolov3) was trained on the localization dataset. 90% of the ResNet classifier knowledge was distilled successfully in the final trained Yolov3 model. The tracker classification accuracy was further improved by proposing a correction methodology that considers both the camera's distance and limited data from the future frames.","abstract_has_math":false,"creators":["Alharin, Alnour"],"institution":"University of Tennessee at Chattanooga","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Sartipi, Mina","Liang, Yu (Hugh); Wu, Dalei","College of Engineering and Computer Science"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2022,"date_issued":"2022-05-31T07:00:00Z","date_published":"2022-05-31T07:00:00Z","updated_at":"2026-07-24T05:47:06Z","subjects":["Computer vision","Machine learning"],"languages":["English","eng"],"rights":[],"rights_urls":["http://rightsstatements.org/vocab/InC/1.0/"],"identifier_entries":[]},"links":{"outbound_url":"https://scholar.utc.edu/theses/698","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Sartipi, Mina","Liang, Yu (Hugh); Wu, Dalei","College of Engineering and Computer Science"]},{"key":"dc:creator","label":"Author","values":["Alharin, Alnour"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2021-05-01T07:00:00Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2022-05-31T07:00:00Z"]},{"key":"dc:publisher","label":"Institution","values":["University of Tennessee at Chattanooga","Chattanooga (Tenn.)"]},{"key":"dc:relation","label":"Dc Relation","values":["Masters Theses and Doctoral Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["Masters theses","Text"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Computer vision","Machine learning"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English","eng"]},{"key":"dc:rights","label":"Dc Rights","values":["http://rightsstatements.org/vocab/InC/1.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://scholar.utc.edu/theses/698"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Dept. of Computational Science","M. S.; A thesis submitted to the faculty of the University of Tennessee at Chattanooga in partial fulfillment of the requirements of the degree of Master of Science."]},{"key":"dc:description.abstract","label":"Abstract","values":["Single-camera multi-vehicle tracking is a preliminary step for traffic optimization. To perform vehicle tracking, a fine-tuned classification dataset was created from multiple sources. A ResNet classifier was trained, and its knowledge was utilized to build a localization dataset automatically without manual annotations. The dataset consists of videos generated from the MLK testbed. Three different object detection frameworks (SSD, Yolov3, and Yolov4) were evaluated with a pipeline that consists of the detector, the fine-tuned classifier and the tracker (DeepSort). Then the detector with the highest accuracy (Yolov3) was trained on the localization dataset. 90% of the ResNet classifier knowledge was distilled successfully in the final trained Yolov3 model. The tracker classification accuracy was further improved by proposing a correction methodology that considers both the camera's distance and limited data from the future frames."]},{"key":"dc:title","label":"Title","values":["Deep learning-based fine-tuned multi-vehicle tracking with classification correction"]}]}],"canonical_facts":{"dc:contributor":["Sartipi, Mina","Liang, Yu (Hugh); Wu, Dalei","College of Engineering and Computer Science"],"dc:creator":["Alharin, Alnour"],"dc:date":["2021-05-01T07:00:00Z"],"dc:date.available":["2022-05-31T07:00:00Z"],"dc:description":["Dept. of Computational Science","M. S.; A thesis submitted to the faculty of the University of Tennessee at Chattanooga in partial fulfillment of the requirements of the degree of Master of Science."],"dc:description.abstract":["Single-camera multi-vehicle tracking is a preliminary step for traffic optimization. To perform vehicle tracking, a fine-tuned classification dataset was created from multiple sources. A ResNet classifier was trained, and its knowledge was utilized to build a localization dataset automatically without manual annotations. The dataset consists of videos generated from the MLK testbed. Three different object detection frameworks (SSD, Yolov3, and Yolov4) were evaluated with a pipeline that consists of the detector, the fine-tuned classifier and the tracker (DeepSort). Then the detector with the highest accuracy (Yolov3) was trained on the localization dataset. 90% of the ResNet classifier knowledge was distilled successfully in the final trained Yolov3 model. The tracker classification accuracy was further improved by proposing a correction methodology that considers both the camera's distance and limited data from the future frames."],"dc:identifier":["https://scholar.utc.edu/theses/698"],"dc:language":["English","eng"],"dc:publisher":["University of Tennessee at Chattanooga","Chattanooga (Tenn.)"],"dc:relation":["Masters Theses and Doctoral Dissertations"],"dc:rights":["http://rightsstatements.org/vocab/InC/1.0/"],"dc:subject":["Computer vision","Machine learning"],"dc:title":["Deep learning-based fine-tuned multi-vehicle tracking with classification correction"],"dc:type":["Masters theses","Text"]},"updated_at":"2026-07-24T05:47:06Z"}