Transfer learning for no-reference image quality metrics using large temporary image sets.
Mykola PonomarenkoSheyda Ghanbaralizadeh BahnemiriKaren O. EgiazarianPublished in: Computational Imaging (2022)
Keyphrases
- transfer learning
- image set
- image quality metrics
- image quality assessment
- quality metrics
- image quality
- quality assessment
- subjective evaluation
- structural similarity
- human visual system
- learning tasks
- reinforcement learning
- quality evaluation
- cross domain
- active learning
- visual quality
- face images
- image matching
- labeled data
- human perception
- multi task
- semi supervised learning
- machine learning
- image database
- data quality
- machine learning algorithms
- text categorization
- text classification
- collaborative filtering
- image data
- learning algorithm
- natural images
- high quality
- video quality
- semi supervised
- text mining
- target domain
- peak signal to noise ratio
- data sources
- quality measures
- training data
- feature extraction
- information retrieval