Study on Memory Color Extraction Method of Image Processing

摘要

A new technique for extract memory colors is proposed. Firstly, color shift caused by different illuminant is corrected. Based on the color characteristics of the image, memory colors could be separated using K-means algorithm while the rest of non-memory colors were filtered off. Finally, to improve the accuracy of memory color extraction, grey world color correction combined with color clustering algorithm were used. Experimental results showed this method can improve the accuracy of memory colors extraction, thus reducing the impact of external light environment when images were captured. It can be implemented as memory colors detection method for many applications.