Total contribution score and fuzzy entropy based two-stage selection of FC, ReLU and inverseReLU features of multiple convolution neural networks for erythrocytes detection.
Sriparna BanerjeeSheli Sinha ChaudhuriPublished in: IET Comput. Vis. (2019)
Keyphrases
- neural network
- fuzzy logic
- detection method
- rich set
- pattern recognition
- false positives
- feature vectors
- low level
- automatic detection
- fuzzy systems
- anomaly detection
- fuzzy sets
- feature set
- classification accuracy
- multiple types
- rule extraction
- adaboost classifier
- membership functions
- feature extraction
- feature subset
- soft computing
- recurrent neural networks
- false alarms
- detection algorithm
- image features
- feature space
- hybrid intelligent
- computer vision