LAUSR

Due to the massive popularity and consciousness towards requirements in evidence, the usage of the surveillance system has tremendously increased. Although video data recorded by the surveillance system contains important information and provides crucial evidence, it is susceptible to malicious alterations. Thus, the authenticity and integrity of the visual evidence need to be examined before the investigation proceeding. In this paper, we propose an integrity verification method for surveillance videos. The proposed method utilizes a randomized hashing method in combination with the elliptic curve cryptography (ECC) for video data integrity verification. In the proposed approach, the video content with a predefined size (segment) is randomized with the unique random value, and then a hash algorithm is applied. The hash algorithm here utilizes the random initialization vector, which is generated with a secret key. Besides, the combination of the randomized hash output and the key is encrypted with the ECC encryption algorithm that ensures the additional security of the data. The experimental results obtained from computer simulation and accident data recorder (ADR)-embedded system show that the proposed method achieves perfect forgery detection for various kinds of tampering such as copy-move, insert, and delete. A complexity analysis based on the execution time for different sized videos shows the minimal overhead of less than 4% for each segment and consumes less memory than the conventional method that utilizes individual frames for hashing.