Mitigating adversarial evasion attacks by deep active learning for medical image classification.
Usman AhmedJerry Chun-Wei LinGautam SrivastavaPublished in: Multim. Tools Appl. (2022)
Keyphrases
- image classification
- active learning
- countermeasures
- feature extraction
- image representation
- bag of words
- information security
- medical diagnosis
- machine learning
- selective sampling
- random sampling
- learning strategies
- medical experts
- semi supervised
- labeled data
- training set
- risk management
- label noise
- visual words
- medical data
- multi label
- malicious attacks
- class specific
- denial of service attacks
- active learning strategies
- learning algorithm
- watermarking scheme
- relevance feedback
- image features
- multi agent
- feature selection
- information systems
- experimental design
- wireless sensor networks
- learning process
- imbalanced data classification