Score thresholding for accurate instance classification in multiple instance learning.
Marc-Andre CarbonneauEric GrangerGhyslain GagnonPublished in: IPTA (2016)
Keyphrases
- multiple instance learning
- image categorization
- supervised learning
- multiple instance
- class labels
- instance selection
- classification accuracy
- machine learning
- support vector machine
- image classification
- feature extraction
- feature space
- semi supervised
- image annotation
- text classification
- support vector
- image segmentation
- decision trees
- data sets
- image retrieval
- training set
- training data
- computer vision
- randomized trees
- object based image retrieval
- positive bags
- diverse density