Color correction on digital image based on reference color charts surrounding object

Abstract

The image color may deviate from the actual colors of the object due to the lighting conditions and different illumination. Therefore, this paper proposes an algorithm for correcting the colors of digital image in various lighting conditions based on matrix calculations from all reference color charts surrounding the object. The six transformation matrix M for color correction have been calculated from six color charts surrounding the object rather than a single color chart. Then the automatic detection and recognition of the target color charts technique has been performed. Next, the image is divided into small grids. The matrix M elements at all grid's corners were calculated based on distance from each color chart position to the considering grid's corner position. Then, the matrix M elements for each pixel in the considering grid were determined by interpolation from the matrix elements of four corners nearby. Finally, the matrix M for each pixel is applied to complete the color correction. This technique has been tested on a table of 132 colors which taken in four different lighting conditions. The results have shown that the color correction technique can be applied for work that requires accurate color matching as a less costly. For our work, this technique will be applied to the teeth photos in order to match the tooth color with the appropriated dental crowns for further development.