於色彩意像空間中進行室內影像之著色

Abstract

Coloring is a frequently used technique in the field of computer graphics. However, the harmony of resultant colors had not yet been evaluated in the conventional algorithms. Previously , a ID image scale of 'CHEERFUL-SILENT' has been constructed to quantitatively evaluate the color harmony of interior images, IN this paper, combined with the evaluation of color-harmony, a coloring system is proposed based on the harmony principles described in the ID color image scale. Our objective is to coloring the interior image according to the grade of human's pleasure/disgust. The procedure is that an interior image is first coded as a color linguistic distribution (CLD), representing mental color impression of the whole image. Furthermore, two types of histogram specification techniques are applied to adaptively modifying the coded CLD by proper references. Based on the resulted CLD, a resultant interior image can be reconstructed by a decoded procedure of color interpolation. The uniform color space of CIE1976 (L*u*v*)is used to preserve the linearity of color additivity. Experimental results demonstrate that the proposed coloring system can obviously improve the mental impression on colors. Also, our potential and further developments are outlined.