Super-resolution processing of the partial pictorial image of the single pictorial image which eliminated artificiality

Abstract

Recently, the number of security cameras (CCTV) for crime prevention has increased rapidly in Japan for the purpose of establishing safe society. As a result, the number of opportunity of utilizing movies of CCTVs has increased in criminal investigation. Last year, we proposed the super-resolution technique using manual sub-pixel shifting process in the 45th ICCST. However, if manually processed images are used as evidence in a criminal trial, a doubt about artificiality in those images might come out. For the images to be treated more properly, a new utilizable super-resolution technique in criminal investigation is proposed in this paper. The biggest point of the proposed technique is that the target images are processed automatically without any artificiality. The new super-resolution technique utilizes a newly designed algorithm with Bilateral Filter. We have developed the new application software using the algorithm. We also compare images processed by the proposed technique with ones processed by the manual sub-pixel shifting process, in order to evaluate the effectiveness of the proposed technique. In conclusion, any person can get the same result by using this technique. Doubts will not arise if images processed by the proposal are used as evidential images in criminal trials.