作者: Mohammad Sulaiman Khorsheed , Ahmad Abdulkader , Hussein Khalid Al-Omari
DOI:
关键词: Character recognition 、 Speech recognition 、 Artificial intelligence 、 Cursive 、 Character (mathematics) 、 Ground truth 、 Segmentation 、 Computer science 、 Natural language processing 、 Word (computer architecture)
摘要: A method and apparatus for automatically identifying character segments recognition is provided. The involves receiving a plurality of words ground truth corresponding to each word the words. may be received in cursive script. Each segmented into one or more based on word. Thereafter, segmentation refined by iteratively re-segmenting similar segments.