Classifier-Guided Visual Correction of Noisy Labels for Image Classification Tasks.
Alex BäuerleHeiko NeumannTimo RopinskiPublished in: Comput. Graph. Forum (2020)
Keyphrases
- image data
- semantic labels
- image classification
- labeled images
- auto annotation
- input image
- training data
- visual appearance
- image content
- single image
- visual features
- visual perception
- multiscale
- image retrieval
- low level
- image representation
- image features
- image regions
- class labels
- test images
- image segmentation
- segmentation method
- label noise
- feature extractor
- image analysis
- classifier training
- visual similarity
- segmented images
- visual information
- similarity measure
- high resolution
- visual data
- feature selection
- feature set
- visual concepts
- web images
- degraded images
- image collections
- segmentation algorithm
- pairwise
- visual effects
- training set
- labeling process
- image processing