作者: Jeong-Hoe Ku
DOI:
关键词:
摘要: Segmentation of a character-containing image, made by placing dark mark against light background, is accomplished establishing vertical pixel projection for each column in the image. For more exact segmentation, image horizontally scanned row after detecting edges characters. The results and thresholds are compared to verify whether spaces noise or discontinuous edges. correlate with neighboring data According comparison, connected eliminated. Thereby, can be segmented into character without segmentation errors.