Deep image hashing based on twin-bottleneck hashing with variational autoencoders.
Maxim VerwilstNina ZizakicLingchen GuAleksandra PizuricaPublished in: MMSP (2021)
Keyphrases
- image segmentation
- image content
- image analysis
- single image
- image data
- template matching
- segmentation method
- image representation
- binary codes
- multiscale
- input image
- low level
- nearest neighbor search
- data structure
- image pixels
- image collections
- image features
- image classification
- vector field
- order preserving
- hash functions
- image structure
- similarity measure
- image matching
- spatial information
- image regions
- keypoints
- image retrieval
- region of interest
- test images
- pixel values
- energy functional
- segmentation algorithm
- denoising
- total variation regularization
- hashing algorithm