Detection of exudates and hemorrhages using an efficient criterion for feature selection.
Dan PopescuLoretta IchimTraian CaramihalePublished in: TSP (2016)
Keyphrases
- diabetic retinopathy
- feature selection
- optic disk
- retinal images
- automatic detection
- early detection
- supervised feature selection
- fundus images
- text categorization
- blood vessels
- mutual information
- detection algorithm
- support vector
- medical experts
- optic disc
- text classification
- feature set
- feature selection algorithms
- selection criterion
- support vector machine
- machine learning
- neyman pearson
- feature space
- classification accuracy
- object detection
- region of interest
- statistical analysis