Learning defect classifiers for visual inspection images by neuro-evolution using weakly labelled training data.
Nils T. SiebelGerald SommerPublished in: IEEE Congress on Evolutionary Computation (2008)
Keyphrases
- visual inspection
- image analysis
- laser scanning
- fluorescence microscopy images
- learning tasks
- neural network
- machine learning
- learning algorithm
- registration accuracy
- object recognition
- labelled training data
- background knowledge
- supervised learning
- active learning
- learning process
- support vector
- training data
- similarity measure
- three dimensional
- pairwise
- image registration
- labeled data
- reinforcement learning
- statistical learning
- decision trees
- feature selection
- data mining