Automatic detection and correction of purple fringing using the gradient information and desaturation

This paper proposes a method to automatically detect and correct purple fringing that is one of the color artifacts due to characteristics of charge coupled device sensors in a digital camera. The proposed method consists of two steps. In the first step, we detect purple fringed regions that satisfy specific properties: hue characteristics around highlight regions with large gradient magnitudes. In the second step, color correction of the purple fringed regions is made by desaturating the pixels in the detected regions. The proposed method is able to detect purple fringe artifacts more precisely than Kang's method. It can be used as a post processing in a digital camera.

[1]  Pietro Fiorentin,et al.  Characterization and calibration of a CCD detector for light engineering , 2005, IEEE Transactions on Instrumentation and Measurement.

[2]  Yap-Peng Tan,et al.  Adaptive color filter array demosaicing with artifact suppression , 2004, 2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512).

[3]  Hsien-Che Lee Introduction to Color Imaging Science: Index , 2005 .

[4]  Takeo Kanade,et al.  The measurement of highlights in color images , 1988, International Journal of Computer Vision.

[5]  Sing Bing Kang Automatic Removal of Chromatic Aberration from a Single Image , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[6]  Xiaolin Wu,et al.  A classification approach to color demosaicking , 2004, 2004 International Conference on Image Processing, 2004. ICIP '04..