Random subspace method for source camera identification

Sensor pattern noise is an inherent fingerprint of imaging devices, which has been widely used for source camera identification, image classification, and forgery detection. In a previous work, we proposed a feature extraction method based on the principal component analysis denoising concept, which can enhance the performance of conventional SPN extraction methods. However, this method is vulnerable, because the training samples are seriously affected by the image content. Accordingly, it is difficult to train a reliable feature extractor by using such a training set. To address this problem, a camera identification framework based on the random subspace method and majority voting is proposed in this work. The experimental results show that the proposed solution can suppress the interference from scene details and enhance the performance in terms of the receiver operating characteristic curve.