Bias-Aware Loss for Training Image and Speech Quality Prediction Models from Multiple Datasets.
Gabriel MittagSaman ZadtootaghajThilo MichaelBabak NaderiSebastian MöllerPublished in: CoRR (2021)
Keyphrases
- quality prediction
- image data
- single image
- random fields
- image content
- image features
- high resolution
- image dataset
- image analysis
- image quality
- image classification
- low level
- statistical model
- image retrieval
- set of training images
- multiscale
- feature extraction
- bayesian framework
- image matching
- parametric models
- image pixels
- input image
- labeled images
- image segmentation
- image representation
- training and testing data
- image collections
- eye fixations
- test images
- high quality
- image regions
- machine learning
- probabilistic model
- structured prediction
- multiresolution
- restricted boltzmann machine
- feature vectors
- training set
- million images
- object recognition
- feature selection