Fingerprinting for scanned comics content identification

In this paper, we propose a fingerprinting to identify comics content using cut information of a comic image. Based the characteristics of a comic image that it consists of several cuts, we design a fingerprinting method which uses the position information of cuts in comic images. Our fingerprinting includes box frame detection, cut information extraction, and cut sequence analyzing. Box frame detection is to detect a normalized box frame from each comic image. Cut information extraction is to extract cut positions using image histogram. We enhanced the identification rate using a sequence of cut positions.