Demosaicing Algorithm for Color Filter Arrays Based on SVMs
One color filter array (CFA) used in a digital camera allows only one of the red-green-blue primary color components to be sensed at each pixel, and interpolating the other missing two components by methods known as demosaicing. A novel support vector machines (SVMs) based demosicing algorithm is proposed to reduce edge artifacts and false color artifacts effectively. The proposed algorithm is a four-step method. Firstly, construct middle plane Kr or Kb on the mosaic image. Secondly, train SVMs with the trained samples constructed on the middle plane. Thirdly, interpolate the unknown value of the middle plane Kr or Kb. Finally, calculate the missing pixel value. Experimental results showed that the proposed approach produced visually pleasing full-color result images and obtained better PSNR values than other demosaicing algorithms.
Keywords: Demosaicing, Color filter array (CFA), Image interpolation, Support vector machines (SVMs).
Download Full-Text
ABOUT THE AUTHORS
Xiao-Fen Jia
Mrs. Jia received the Master degree in control science and engineering, from the Harbin Institute of Technology. Currently, she is a lectorate at Anhui University of Science & Technology, Electrical and Information Engineering College. Her research interests include Image processing and Rough sets.
Bai-Ting Zhao
Dr. Zhao received the Master degree in control theory and control engineering from the Qingdao University of Science & Technology, in 2005. He received the Ph.D. degree in control science and engineering, from the Harbin Institute of Technology. Currently, he is a lectorate at Anhui University of Science & Technology, Electrical and Information Engineering College. His research interests include Image processing, intelligent control and Rough sets.
Xiao-Fen Jia
Mrs. Jia received the Master degree in control science and engineering, from the Harbin Institute of Technology. Currently, she is a lectorate at Anhui University of Science & Technology, Electrical and Information Engineering College. Her research interests include Image processing and Rough sets.
Bai-Ting Zhao
Dr. Zhao received the Master degree in control theory and control engineering from the Qingdao University of Science & Technology, in 2005. He received the Ph.D. degree in control science and engineering, from the Harbin Institute of Technology. Currently, he is a lectorate at Anhui University of Science & Technology, Electrical and Information Engineering College. His research interests include Image processing, intelligent control and Rough sets.
