Concept-oriented labelling of patent images based on Random Forests and proximity-driven generation of synthetic data.
Dimitris LiparasAnastasia MoumtzidouStefanos VrochidisIoannis KompatsiarisPublished in: VL@COLING (2014)
Keyphrases
- synthetic data
- random forests
- real image data
- input image
- random forest
- logistic regression
- depth images
- object recognition
- data sets
- information retrieval
- image features
- image classification
- decision trees
- text categorization
- feature points
- pairwise
- image annotation
- similarity measure
- computer vision
- machine learning