DeVIS: Making Deformable Transformers Work for Video Instance Segmentation

Video Instance Segmentation (VIS) jointly tackles multi-object detection, tracking, and segmentation in video sequences. In the past, VIS methods mirrored the fragmentation of these subtasks in their architectural design, hence missing out on a joint solution. Transformers recently allowed to cast t...

Full description

Bibliographic Details
Author: Caelles Prat, Adrià
Format: master thesis
Publication Date:2022
Country:España
Institution:Universitat Politècnica de Catalunya (UPC)
Repository:UPCommons. Portal del coneixement obert de la UPC
Language:English
OAI Identifier:oai:upcommons.upc.edu:2117/382651
Online Access:https://hdl.handle.net/2117/382651
Access Level:Open access
Keyword:Video recording
Image segmentation
Deformable Transformers
Video Instance Segmentation
Vídeo
Imatges--Segmentació
Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Processament del senyal::Processament de la imatge i del senyal vídeo
Description
Summary:Video Instance Segmentation (VIS) jointly tackles multi-object detection, tracking, and segmentation in video sequences. In the past, VIS methods mirrored the fragmentation of these subtasks in their architectural design, hence missing out on a joint solution. Transformers recently allowed to cast the entire VIS task as a single set-prediction problem. Nevertheless, the quadratic complexity of existing Transformer-based VIS methods requires long training times, high memory requirements, and processing of low-single-scale feature maps.Deformable attention provides a more efficient alternative but its application to the temporal domain or the segmentation task have not yet been explored. In this work, we present Deformable VIS (DeVIS), a VIS method which capitalizes on the efficiency and performance of deformable Transformers. To reason about all VIS subtasks jointly over multiple frames, we present temporal multi-scale deformable attention with instance-aware object queries. We further introduce a new image and video instance mask head which exploits multi-scale features, and perform near-online video processing with multi-cue clip tracking. DeVIS benefits from comparatively small memory as well as training time requirements, and achieves state-of-the-art results on the YouTube-VIS 2019 and 2021, as well as the challenging OVIS dataset.