Sorry, you need to enable JavaScript to visit this website.

TokenMotion: Motion-Guided Vision Transformer For Video Camouflaged Object Detection Via Learnable Token Selection

DOI:
10.60864/jppb-e006
Citation Author(s):
Zifan Yu, Erfan Bank Tavakoli, Meida Chen, Suya You, Raghuveer Rao, Sanjeev Agarwal, and Fengbo Ren
Submitted by:
Zifan Yu
Last updated:
6 June 2024 - 10:33am
Document Type:
Presentation Slides
Document Year:
2024
Event:
Presenters:
Raghuveer Rao
Paper Code:
IVMSP-L10.2
 

The area of Video Camouflaged Object Detection (VCOD) presents unique challenges in the field of computer vision due to texture similarities between target objects and their surroundings, as well as irregular motion patterns caused by both objects and camera movement. In this paper, we introduce TokenMotion (TMNet), which employs a transformer-based model to enhance VCOD by extracting motion-guided features using a learnable token selection. Evaluated on the challenging MoCA-Mask dataset, TMNet achieves state-of-the-art performance in VCOD. It outperforms the existing state-of-the-art method by a 1.7\% improvement in weighted F-measure, an 1.7% enhancement in S-measure, and a 2.1\% boost in mean Dice. The results demonstrate the benefits of utilizing motion-guided features via learnable token selection within a transformer-based framework to tackle the intricate task of VCOD. The code of our work will be available when the paper is accepted.

up
0 users have voted: