Digital image forgery detection through data embedding in spatial domain and cellular automata

In this paper we present a new algorithm to detect digital image forgery based on cellular automata and data embedding in spatial domain. The original JPEG image will be partitioned into some regions. We use region-based segmentation to specifying the desired regions of interest from the input image. First we extract the visual attributes of the original image and achieve the statistical information for the selected region. Then we apply linear cellular automata rules to create a robust cipher key from these values. We embed the cipher key into the spatial domain to authenticate and validate the original image. The proposed algorithm is applied on 100 numbers of grayscale images (size 800 × 600). The results have demonstrated the robustness and stable time complexity of the proposed method.