A bottom-up method using texture features and a graph-based representation for lettrine recognition and classification.
Maroua MehriPetra Gomez-KrämerPierre HérouxMickaël CoustatyJulien LerougeRémy MullotPublished in: ICDAR (2015)
Keyphrases
- classification accuracy
- texture features
- feature set
- feature extraction
- texture feature extraction
- global image features
- object recognition
- feature vectors
- machine learning
- recognition algorithm
- similarity measure
- image features
- gray level
- multiresolution
- texture analysis
- texture classification
- decision trees
- data sets