DC Field | Value | Language |
---|---|---|
dc.contributor.author | Im, Woobin | ko |
dc.contributor.author | Lee, Sebin | ko |
dc.contributor.author | Yoon, Sung-eui | ko |
dc.date.accessioned | 2022-11-01T10:00:14Z | - |
dc.date.available | 2022-11-01T10:00:14Z | - |
dc.date.created | 2022-11-01 | - |
dc.date.created | 2022-11-01 | - |
dc.date.issued | 2022-10-25 | - |
dc.identifier.citation | 2022 European Conference on Computer Vision (ECCV), pp.302 - 318 | - |
dc.identifier.issn | 0302-9743 | - |
dc.identifier.uri | http://hdl.handle.net/10203/299200 | - |
dc.description.abstract | A training pipeline for optical flow CNNs consists of a pretraining stage on a synthetic dataset followed by a fine tuning stage on a target dataset. However, obtaining ground truth flows from a target video requires a tremendous effort. This paper proposes a practical fine tuning method to adapt a pretrained model to a target dataset without ground truth flows, which has not been explored extensively. Specifically, we propose a flow supervisor for self-supervision, which consists of parameter separation and a student output connection. This design is aimed at stable convergence and better accuracy over conventional self-supervision methods which are unstable on the fine tuning task. Experimental results show the effectiveness of our method compared to different self-supervision methods for semi-supervised learning. In addition, we achieve meaningful improvements over state-of-the-art optical flow models on Sintel and KITTI benchmarks by exploiting additional unlabeled datasets. Code is available at https://github.com/iwbn/flow-supervisor. | - |
dc.language | English | - |
dc.publisher | IEEE Computer Society and the Computer Vision Foundation (CVF) | - |
dc.title | Semi-Supervised Learning of Optical Flow by Flow Supervisor | - |
dc.type | Conference | - |
dc.identifier.wosid | 000903538700018 | - |
dc.identifier.scopusid | 2-s2.0-85144505790 | - |
dc.type.rims | CONF | - |
dc.citation.beginningpage | 302 | - |
dc.citation.endingpage | 318 | - |
dc.citation.publicationname | 2022 European Conference on Computer Vision (ECCV) | - |
dc.identifier.conferencecountry | IS | - |
dc.identifier.conferencelocation | Tel Aviv | - |
dc.identifier.doi | 10.1007/978-3-031-19833-5_18 | - |
dc.contributor.localauthor | Yoon, Sung-eui | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.