A novel video sign image expander

Abstract

A simple hardware for expanding sign images in a video program is constructed for aiding deaf TV viewers. It is designed as a TV set-top box in which the composite video is digitized, processed and encoded back into an analog form for the TV video input. The image expansion is implemented using a medium size FPGA and two banks of 32 Kbytes video memories. The sign image can be as large as 128/spl times/128 pixels or approximately 1/25 of a TV screen. It is captured from the digitized video signal and stored in one memory bank waiting to be expanded and inserted back at the next video frame. The expanded sign information will be delayed by one video frame or 1/25 sec which should not be noticeable. Using an infrared remote controller, a viewer can choose 9 different image scaling ratios from 1/spl times/1 to 2/spl times/2. The FPGA synthesizes only a fixed 2/spl times/2 expanding circuit using bilinear interpolation and implements a nearest neighborhood technique to get any other reduced ratio from the 2/spl times/2 expanded image. This results in minimal hardware and low cost. A small 8 bit microcontroller is also implemented in the FPGA to handle infrared control commands, switch the memory banks and program the scaling ratio, location and size of the sign window.