Abstract: A new computer\|aided analogy approach for color restoration of ancient mural is proposed in this paper. The mural images whose colors are relatively intact are stored, along with their feature parameters used to describe the murals, into the database as the analogy source for color restoration of faded murals. The feature parameters are classified into three levels according to different factors that determine and affect the color of mural. The database can be retrieved from one or more than one level. In order to achieve a uniform perception with human vision, SHI color space is chosen. Several techniques for color image processing are used to process the mural image. These include color object extraction based on interactive region growing and color transformation based on color histogram.