{"id":{"repo_id":"mit","oai_identifier":"oai:dspace.mit.edu:1721.1/151425"},"canonical_url":"https://search.dev.ndltd.org/etd/mit/oai:dspace.mit.edu:1721.1/151425","repository":{"repo_id":"mit","name":"MIT","base_url":"https://dspace.mit.edu/oai/request"},"display":{"title":"Learning Privacy-Preserving Transferable Video Representations","abstract":"Pretraining on massive video datasets has become essential to achieve high action recognition performance on smaller downstream datasets. However, most large-scale video datasets are accompanied with issues related to privacy, ethics, and data protec-tion, often preventing them to be publicly shared with the community for reproducible research. Existing work has attempted to alleviate these problems by blurring faces, downsampling videos, or training on synthetic data. On the other hand, analysis on the transferability of privacy-preserving pretrained models to downstream tasks has been limited. In this work, we study this problem by first asking the question: can we pretrain models for human action recognition with data that does not include humans? To this end, we present, for the first time, a benchmark that leverages real-world videos with humans removed and synthetic data containing virtual humans to pretrain a model. We then evaluate the transferability of the representation learned on this data to a diverse set of downstream action recognition datasets. Furthermore, we propose a novel pre-training strategy, called Privacy-Preserving MAE-Align, to effectively combine synthetic data and human-removed real data. Compared to previous baselines, our approach reduces, by a large margin, the performance gap between human and no-human action recognition representations on downstream tasks. Our benchmark, code, and models will be made publicly available.","abstract_html":"Pretraining on massive video datasets has become essential to achieve high action recognition performance on smaller downstream datasets. However, most large-scale video datasets are accompanied with issues related to privacy, ethics, and data protec-tion, often preventing them to be publicly shared with the community for reproducible research. Existing work has attempted to alleviate these problems by blurring faces, downsampling videos, or training on synthetic data. On the other hand, analysis on the transferability of privacy-preserving pretrained models to downstream tasks has been limited. In this work, we study this problem by first asking the question: can we pretrain models for human action recognition with data that does not include humans? To this end, we present, for the first time, a benchmark that leverages real-world videos with humans removed and synthetic data containing virtual humans to pretrain a model. We then evaluate the transferability of the representation learned on this data to a diverse set of downstream action recognition datasets. Furthermore, we propose a novel pre-training strategy, called Privacy-Preserving MAE-Align, to effectively combine synthetic data and human-removed real data. Compared to previous baselines, our approach reduces, by a large margin, the performance gap between human and no-human action recognition representations on downstream tasks. Our benchmark, code, and models will be made publicly available.","abstract_has_math":false,"creators":["Zhong, Howard"],"institution":"Massachusetts Institute of Technology","degree_name":"Master","degree_level":null,"degree_discipline":null,"degree_department":"Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science","school":null,"contributors":[],"advisors":["Oliva, Aude"],"committee_chairs":[],"committee_members":[],"year":2023,"date_issued":"2023-06","date_published":"2023-06","updated_at":"2026-07-22T22:21:07Z","subjects":[],"languages":[],"rights":["In Copyright - Educational Use Permitted","Copyright retained by author(s)"],"rights_urls":["https://rightsstatements.org/page/InC-EDU/1.0/"],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/1721.1/151425","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Oliva, Aude"]},{"key":"dc:contributor.department","label":"Department","values":["Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science"]},{"key":"dc:creator","label":"Author","values":["Zhong, Howard"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2023-07-31T19:38:47Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2023-07-31T19:38:47Z"]},{"key":"dc:date.issued","label":"Date","values":["2023-06"]},{"key":"dc:publisher","label":"Institution","values":["Massachusetts Institute of Technology"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master","Master of Engineering in Electrical Engineering and Computer Science"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:rights","label":"Dc Rights","values":["In Copyright - Educational Use Permitted","Copyright retained by author(s)"]},{"key":"dc:rights.uri","label":"Rights URI","values":["https://rightsstatements.org/page/InC-EDU/1.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/1721.1/151425"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Pretraining on massive video datasets has become essential to achieve high action recognition performance on smaller downstream datasets. However, most large-scale video datasets are accompanied with issues related to privacy, ethics, and data protec-tion, often preventing them to be publicly shared with the community for reproducible research. Existing work has attempted to alleviate these problems by blurring faces, downsampling videos, or training on synthetic data. On the other hand, analysis on the transferability of privacy-preserving pretrained models to downstream tasks has been limited. In this work, we study this problem by first asking the question: can we pretrain models for human action recognition with data that does not include humans? To this end, we present, for the first time, a benchmark that leverages real-world videos with humans removed and synthetic data containing virtual humans to pretrain a model. We then evaluate the transferability of the representation learned on this data to a diverse set of downstream action recognition datasets. Furthermore, we propose a novel pre-training strategy, called Privacy-Preserving MAE-Align, to effectively combine synthetic data and human-removed real data. Compared to previous baselines, our approach reduces, by a large margin, the performance gap between human and no-human action recognition representations on downstream tasks. Our benchmark, code, and models will be made publicly available."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["M.Eng."]},{"key":"dc:title","label":"Title","values":["Learning Privacy-Preserving Transferable Video Representations"]}]}],"canonical_facts":{"dc:contributor.advisor":["Oliva, Aude"],"dc:contributor.department":["Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science"],"dc:creator":["Zhong, Howard"],"dc:date.accessioned":["2023-07-31T19:38:47Z"],"dc:date.available":["2023-07-31T19:38:47Z"],"dc:date.issued":["2023-06"],"dc:description.abstract":["Pretraining on massive video datasets has become essential to achieve high action recognition performance on smaller downstream datasets. However, most large-scale video datasets are accompanied with issues related to privacy, ethics, and data protec-tion, often preventing them to be publicly shared with the community for reproducible research. Existing work has attempted to alleviate these problems by blurring faces, downsampling videos, or training on synthetic data. On the other hand, analysis on the transferability of privacy-preserving pretrained models to downstream tasks has been limited. In this work, we study this problem by first asking the question: can we pretrain models for human action recognition with data that does not include humans? To this end, we present, for the first time, a benchmark that leverages real-world videos with humans removed and synthetic data containing virtual humans to pretrain a model. We then evaluate the transferability of the representation learned on this data to a diverse set of downstream action recognition datasets. Furthermore, we propose a novel pre-training strategy, called Privacy-Preserving MAE-Align, to effectively combine synthetic data and human-removed real data. Compared to previous baselines, our approach reduces, by a large margin, the performance gap between human and no-human action recognition representations on downstream tasks. Our benchmark, code, and models will be made publicly available."],"dc:description.degree":["M.Eng."],"dc:identifier.uri":["https://hdl.handle.net/1721.1/151425"],"dc:publisher":["Massachusetts Institute of Technology"],"dc:rights":["In Copyright - Educational Use Permitted","Copyright retained by author(s)"],"dc:rights.uri":["https://rightsstatements.org/page/InC-EDU/1.0/"],"dc:title":["Learning Privacy-Preserving Transferable Video Representations"],"dc:type":["Thesis"],"thesis:degree_name":["Master","Master of Engineering in Electrical Engineering and Computer Science"]},"updated_at":"2026-07-22T22:21:07Z"}