Implementation of Video Authentication using Sensor Pattern Noise

Abstract: Video evidences play an important role in criminal investigations for obtaining detailed information from their own. To deduce whether the contents of the given video are authentic or exactly similar to the captured one is determined using video authentication technique. Video authentication techniques are used for detecting malicious tampering and preventing various types of forgeries, performed on video data. Pattern noises acquired are treated as the fingerprint of devices which is very useful in for device identification. Image/video source identification is a difficult task on typical embedded devices. The paper provides a hardware or FPGA implementation of Modified filter bank (FB) based orthogonal db2 DWT filter and an image denoising algorithm based on minimum mean square error estimation. Without decrease in authentication accuracy, there is saving in hardware and improvement in clock frequency.

Keywords: Video authentication; Sensor Pattern Noise;DWT; Denoising.