Video reduction method based subtitle caption detection and recognition

The present invention is a method of reducing video caption based on the caption detection and recognition, comprising the following steps: Step 1: The video uploaded to the analysis apparatus; Step 2: detecting a subtitle region on each frame; Step 3: detecting by a sliding window caption; step 4: the subtitle detected, wherein the certain frame, using the OCR technique, characters recognized, starting with the subtitle text and deadline of storing the text subtitle. The present invention is detected by the subtitle, the subtitle detected video hard, and optical character recognition technology to recognize that the content of hard subtitles, soft reduction to give the subtitle (text subtitle).