Bridging the Gap between Object and Image-level Representations for Open-Vocabulary Detection.
Hanoona Abdul RasheedMuhammad MaazMuhammad Uzair KhattakSalman H. KhanFahad Shahbaz KhanPublished in: NeurIPS (2022)
Keyphrases
- bounding box
- pixel level
- image data
- single image
- image content
- higher level
- partial occlusion
- image segmentation
- image regions
- input image
- image features
- target object
- cluttered scenes
- object shapes
- keypoints
- image collections
- lighting conditions
- region of interest
- multiscale
- low level
- cluttered background
- background clutter
- similar objects
- object models
- spatial relationships
- object model
- multiple objects
- object representation
- object level
- image analysis
- ground plane
- image retrieval
- edge detection
- object recognition
- object detection
- feature points
- test images
- image representation
- spatial information
- three dimensional objects
- hough transform
- normalized correlation
- object class
- detecting objects
- image segments
- object shape
- perceptual grouping
- segmentation method
- high resolution
- d objects
- object categories
- detection method