Amicable Aid: Perturbing Images to Improve Classification Performance.
Juyeop KimJun-Ho ChoiSoobeom JangJong-Seok LeePublished in: ICASSP (2023)
Keyphrases
- image classification
- image data
- input image
- pattern recognition
- image analysis
- image database
- ground truth
- image retrieval
- machine learning
- extracting features
- support vector machine svm
- classification accuracy
- image collections
- feature vectors
- correct classification rate
- satellite images
- image registration
- feature extraction
- three dimensional
- test images
- high resolution computed tomography
- region of interest
- classification method
- object recognition
- text classification
- image matching
- decision trees
- feature representation
- fractal dimension
- image features
- edge detection
- class labels
- gabor filters
- image annotation
- lighting conditions
- hyperspectral images
- skin lesion
- preprocessing stage
- similarity measure
- image content