Identification and Extraction of Text in Non- Blurred and Blurred Image Sequences
Sudharshan P. Duth, H.L. Rahul · 2020
The images embedded with text contain information and therefore it is important to identify and extract text. Extracting text in a blurred image is a challenging task due to movement and defocus, whereas it is an easy task in the nonblurred image. In this paper, text identification from non-blur and blur images is done using an Otsu method and region of interest (ROI) and text are extracted using a convolution neural network (CNN). The evaluation results show 98.1% accuracy for non-blurred images and 96% accuracy for blurred images. The text is extracted and identified with efficient accuracies than the existing methods.