Examining convolutional feature extraction using Maximum Entropy (ME) and Signal-to-Noise Ratio (SNR) for image classification.
Nidhi GowdraRoopak SinhaStephen G. MacDonellPublished in: CoRR (2021)
Keyphrases
- signal to noise ratio
- maximum entropy
- image classification
- feature extraction
- sparse coding
- noise reduction
- maximum entropy principle
- feature vectors
- principle of maximum entropy
- image processing
- edge detection
- wavelet transform
- matched filter
- sparse representation
- minimum cross entropy
- conditional random fields
- transformation based learning
- bit error rate
- class specific
- gaussian noise
- feature space
- principal component analysis
- image features
- minimum mean square error
- face recognition
- frequency domain
- wiener filter
- received signal
- iterative scaling
- feature selection