earticle

영어

Existing code book algorithm removes the background or the specified image area by composing a code Book from the basis image and mapping it on an objective image to make a code book. The code book algorithm is known to use the color similarity such that it removes by clustering like colors, however, it is disadvantageous that larger the code size or the objective image, slower the removal process. To improve the situation, this study suggests the variable order n×m dimensional vector. The suggested algorithm generates the variable order vectors with received values of n and m, calculates the average of 3 pixels around the objective image and replaces each pixel. The replaced pixels reduces the image by dividing it into squares implementing the code book and removes the background color by applying the vectors again. Then the image is finally enlarged to the original size using the linear interpolation. Since the vectors are applied to the reduced objective image to remove the background in this algorithm, the speed which has been the shortcoming of the code book is improved. It is confirmed that the removal of the background in complex color images where the background was difficult to distinguish is successful as a result of application of the proposed algorithm to the actual objective image.