3D-aware instance segmentation and tracking in egocentric videos

Conference item

Egocentric videos present unique challenges for 3D scene understanding due to rapid camera motion, frequent object occlusions, and limited object visibility. This paper introduces a novel approach to instance segmentation and tracking in first-person video that leverages 3D awareness to overcome these obstacles. Our method integrates scene geometry, 3D object centroid tracking, and instance segmentation to create a robust framework for analyzing dynamic egocentric scenes. By incorporating spatial and temporal cues, we achieve superior performance compared to state-of-the-art 2D approaches. Extensive evaluations on the challenging EPIC Fields dataset demonstrate significant improvements across a range of tracking and segmentation consistency metrics. Specifically, our method outperforms the next best performing approach by 7 points in Association Accuracy (AssA) and 4.5 points in IDF1 score, while reducing the number of ID switches by 73% to 80% across various object categories. Leveraging our tracked instance segmentations, we showcase downstream applications in 3D object reconstruction and amodal video object segmentation in these egocentric settings.

Authors: Bhalgat, Y; Tschernezki, V; Laina, I; Henriques, JF; Vedaldi, A

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Springer
• Host title: Computer Vision – ACCV 2024
• Pages: 347-364
• Series: Lecture Notes in Computer Science
• Series number: 15474
• Publication date: 2024-12-07
• Acceptance date: 2024-09-20
• Event title: 17th Asian Conference on Computer Vision (ACCV 2024)
• Event location: Hanoi, Vietnam
• Event website: https://accv2024.org/
• Event start date: 2024-12-08
• Event end date: 2024-12-12
• DOI: 10.1007/978-981-96-0908-6_20
• EISSN: 1611-3349
• ISSN: 0302-9743
• EISBN: 9789819609086
• ISBN: 9789819609079
• Language: English
• Keywords: 3D aware tracking; video object segmentation; egocentric understanding