PRNU Extraction from Stabilized Video: a Patch Maybe Better than a Bunch

This paper presents an algorithm to solve the problem of Photo-Response Non-Uniformity (PRNU) noise facing stabilized video. The video undergoes in-camera processing like rolling shutter correction. Thus, misalignment exists between PRNU noises in adjacent frames owing global and local frame registration performed by processing. makes reference test unable extract match accurately. We design a computing method maximum likelihood estimation for extracting from frames. Besides, unlike most prior arts tending whole frame, we propose new patch-based matching strategy, aiming at reducing influence noise. After noise, this adopts overlapping matching. It is different traditional method. conducts experiments on 224 videos taken 13 smartphones VISION database. area under curve proposed 0.841, which significantly higher than 0.805 algorithm. Experimental results show good performance effectiveness strategy comparing with arts.