A2-LINK: Recognizing Disguised Faces via Active Learning and Adversarial Noise Based Inter-Domain Knowledge.
Anshuman SuriMayank VatsaRicha SinghPublished in: IEEE Trans. Biom. Behav. Identity Sci. (2020)
Keyphrases
- domain knowledge
- active learning
- label noise
- selective sampling
- relevance feedback
- noise level
- learning algorithm
- random noise
- face recognition
- noise reduction
- semi supervised
- background knowledge
- imbalanced data classification
- additive noise
- domain experts
- learning strategies
- training examples
- machine learning
- computer vision
- noisy environments
- image noise
- multi agent
- prior knowledge
- signal to noise ratio
- noisy data
- supervised learning
- labeled data
- knowledge sources
- face images
- feature selection
- batch mode
- knowledge base
- noise free
- data sets
- arbitrary shape
- training data
- link structure
- gaussian noise
- class imbalance
- random sampling