Optimal and simultaneous designs of Hermitian transforms and masks for reducing intraclass separations of feature vectors for anomaly detection of diabetic retinopathy images.
Suba Raman SubramaniamApostolos GeorgakisBingo Wing-Kuen LingJonathan GohHongying Lilian TangTünde PetöGeorge SalehPublished in: CSNDSP (2012)
Keyphrases
- anomaly detection
- feature vectors
- image data
- intra class
- intrusion detection
- edge detection
- image database
- image features
- image analysis
- image retrieval
- input image
- intrusion detection system
- object recognition
- diabetic retinopathy
- similarity measure
- test images
- image registration
- image classification
- support vector machine
- face images
- image processing
- machine learning
- co occurrence
- feature space
- distance measure
- maximum likelihood
- decision trees
- data sets