ComplETR: Reducing the cost of annotations for object detection in dense scenes with vision transformers.
Achin JainKibok LeeGurumurthy SwaminathanHao YangBernt SchieleAvinash RavichandranOnkar DabeerPublished in: CoRR (2022)
Keyphrases
- object detection
- computer vision
- object categories
- vision system
- real time
- minimum cost
- pedestrian detection
- high cost
- image understanding
- semantic annotation
- image annotation
- object segmentation
- face detection
- moving objects
- real scenes
- metadata
- scene recognition
- scene understanding
- deformable part models
- total cost
- dynamic scenes
- natural scenes
- cost sensitive
- image processing
- neural network