A Novel Dataset for English-Arabic Scene Text Recognition (EASTR)-42K and Its Evaluation Using Invariant Feature Extraction on Detected Extremal Regions.
Saad Bin AhmedSaeeda NazMuhammad Imran RazzakRubiyah Bte YusofPublished in: IEEE Access (2019)
Keyphrases
- feature extraction
- scene text recognition
- arabic language
- feature set
- natural language
- preprocessing
- image processing
- feature vectors
- maximally stable
- principal component analysis
- frequency domain
- cross language
- graph theory
- region descriptors
- region detection
- planar objects
- language identification
- morphological analysis
- english language
- handwriting recognition
- texture analysis
- image regions
- keypoints
- language model
- input image
- object recognition