Setting our target to development of new hardware, we have been studying a technique of enhancing the power of gradation expression, which will require less amount of calculation and will be unlikely to decrease the sharpness of image. With our conventional system, false detection of contour areas was observed and correction of gradation is not sufficient in some cases if noise is superposed on the video picture signals. Hence, in this study, we have proposed a new algorithm for improving the method of contour detection. After evaluation by image simulation, we have evaluated the performance of hardware employing this algorithm. As the result, in spite of less amount of calculation, we have confirmed that gradation can be corrected effectively even over video picture signals on which noise is superposed.