Alleviating Noisy-label Effects in Image Classification via Probability Transition Matrix.
Ziqi ZhangYuexiang LiHongxin WeiKai MaTao XuYefeng ZhengPublished in: BMVC (2021)
Keyphrases
- image classification
- transition matrix
- multi label
- markov chain
- bag of words
- probability distribution
- feature extraction
- image representation
- label noise
- transition probabilities
- image features
- class membership
- finite state
- visual words
- directed graph
- text classification
- incomplete data
- multi layer
- feature selection
- noisy environments
- computer vision
- machine learning