Tokenmotion: Motion-Guided Vision Transformer for Video Camouflaged Object Detection VIA Learnable Token Selection.
Zifan YuErfan Bank TavakoliMeida ChenSuya YouRaghuveer RaoSanjeev AgarwalFengbo RenPublished in: ICASSP (2024)
Keyphrases
- object detection
- space time
- computer vision
- object motion
- temporal filtering
- video sequences
- key frames
- real time
- motion estimation
- video data
- spatial and temporal
- visual perception
- input video
- successive frames
- moving camera
- motion features
- static images
- dynamic scenes
- video content
- motion analysis
- camera motion
- video frames
- video streams
- visual cues
- low frame rate
- vision system
- visual data
- motion capture data
- optical flow
- video objects
- face detection
- background subtraction
- shot change detection
- image sequences
- temporal continuity
- object segmentation
- moving objects
- motion model
- motion patterns
- visual input
- multi class
- temporal consistency
- video scene
- video summarization
- motion trajectories
- human motion
- motion field
- layered representation
- human actions
- temporal coherence
- dynamic textures
- video analysis
- wide area motion imagery
- action detection
- surveillance videos
- scene understanding
- multimedia
- motion detection
- motion segmentation
- visual information
- fault diagnosis
- learning algorithm