(ASNA) An Attention-Based Siamese-Difference Neural Network With Surrogate Ranking Loss Function for Perceptual Image Quality Assessment.
Seyed Mehdi AyyoubzadehAli RoyatPublished in: CVPR Workshops (2021)
Keyphrases
- loss function
- image quality assessment
- learning to rank
- neural network
- human visual system
- perceptual quality
- quality assessment
- image quality
- structural similarity
- pairwise
- image database
- human perception
- visual perception
- support vector
- natural images
- correlation coefficient
- reduced reference
- quality evaluation
- quality metrics
- saliency map
- visual attention
- ranking functions
- image quality metrics
- back propagation
- natural scenes
- kullback leibler divergence
- evaluation measures
- perceptual image quality
- quality measures
- ranking algorithm
- web search
- evaluation metrics
- visual quality
- visual information
- information retrieval
- information theoretic
- bit rate
- semi supervised
- low level
- high quality
- machine learning