Image Captioning with Unseen Objects.
Berkan DemirelRamazan Gokberk CinbisNazli Ikizler-CinbisPublished in: BMVC (2019)
Keyphrases
- image pixels
- input image
- multiple objects
- multiscale
- image segmentation
- image regions
- image features
- complex scenes
- image data
- spatial relations
- object features
- image content
- image analysis
- bounding box
- hough transform
- keypoints
- image representation
- spatial relationships
- high resolution
- single image
- image retrieval
- image matching
- image classification
- feature values
- test images
- visual appearance
- static images
- image processing operations
- pixel values
- similar objects
- geometric constraints
- edge detection
- real world scenes
- foreground background separation
- previously unseen
- real world objects
- sample images
- image segments
- object models
- target object
- partial occlusion
- image collections
- template matching
- object recognition
- relative position
- aerial images
- three dimensional objects
- object model
- object segmentation
- lighting conditions
- spatial information
- low level
- image sequences