In this paper, we describe a signal processing method for upgrading an image quality based on a reduction of contour artifacts. In order to smooth the contour out, some methods have been reported in the image restoration field that had been averaging a pixel code value around the contour. However, these methods have some disadvantages such as hardware complexity and resolution loss. On the contrary, the proposed signal processing method makes it possible to reduce contours without any resolution loss by smaller hardware such as block-based LUTs (look-up tables), since a bit depth of input signal is expanded based on a spatial frequency distribution of the input signal and on the property of the human visual system. By the above features, this method is especially effective, when the bit depth of an input signal is low compared with a bit depth of a display device, or when the display capability of a digital display device improves in the future and became to be able to display more bits than the input signal.