A STUDY ON SKIN TONE BASED ROI SEGMENTATION TO EXTRACT NEWSREADER REGION FROM HEARING IMPAIRED NEWS VIDEOS

In this research article, the video segmentation based on skin color using Region of Interest is presented. The objective of the research is capturing the sign words from the deaf and dumb news videos and preparing textual information. By default, video processing requires huge memory space to process the frames. The time taken for processing frames and pixels are also too high. For this research, the news video has been pre-processed and presented for object recognition process. The skin tone of the news reader varies with respective to the attributes of the images like brightness, luminescence and the color of the video. When broadcasting an edited-video, there are possibilities in its resolution. The resolution also affects the skin tone of the news reader. Hence skin tone selection based region of interest for news-video is analyzed and presented to gain knowledge for video segmentation