VEIL: Vetting Extracted Image Labels from In-the-Wild Captions for Weakly-Supervised Object Detection.
Arushi RaiAdriana KovashkaPublished in: EACL (1) (2024)
Keyphrases
- weakly supervised
- semantic segmentation
- weakly labeled
- object detection
- object detectors
- object class
- image data
- input image
- image features
- fully supervised
- superpixels
- multiscale
- image classification
- image content
- test images
- image segmentation
- object categories
- single image
- spatial information
- image retrieval
- conditional random fields
- bounding box
- region of interest
- partial occlusion
- object classes
- segmentation algorithm
- multi class
- high resolution
- object recognition
- computer vision