Word Detection Based on unet model

Since computer vision developed drastically, OCR, one of the read of computer vision also developed rapidly. However, most of the OCR papaers are focused on finding word and recognizing it by image file. Nowdays, many videos uploaded in youtube and we can get a lot of data from it. So I try to apply OCR to videos with accumulation of frames and unet based model.

This is a sample result with finding area of words by using own model.

You can get more information about it by downloading my paper