Label-Attention Transformer with Geometrically Coherent Objects for Image Captioning.
Shikha DubeyFarrukh OlimovMuhammad Aasim RafiqueJoonmo KimMoongu JeonPublished in: CoRR (2021)
Keyphrases
- image pixels
- bounding box
- image data
- image classification
- low level
- partial occlusion
- image regions
- image analysis
- multiple objects
- image features
- image content
- complex scenes
- single image
- multiscale
- keypoints
- visual appearance
- real world objects
- target object
- image segments
- segmentation algorithm
- lighting conditions
- real world scenes
- high resolution
- relative position
- input image
- image retrieval
- individual objects
- similar objects
- pixel values
- image segmentation
- similarity measure
- fuzzy logic
- image representation
- object detection
- edge detection
- feature points
- computer vision
- object features
- moving objects
- closed curves
- image collections
- object models
- segmentation method
- image matching
- region of interest
- spatial relations
- object segmentation