Integration Processing of Extraction of Telop Regions from VideoImages and Application of OCR

摘要

The contents information from video image is included in the telop. Each telop area of video frame image represents an area composed of character patterns. The high performance OCR seems a quite useful software to recognize them and to produce the contents description by symbols. Prom viewpoint of OCR, it has a problem that it requires more precise detction of character area. The character area detction from an video frame is a quite difficult problem. This problem becomes easy to solve if OCR can allow the rough detction of telop area. Integration processing of OCR application and telop area detection should be realized to solve the problems existing both sides. In this paper, we proposes a framework toward the integration.