Mahmoud R. El-Sakka and Mohamed S. Kamel, "Detection Of Image Format Using Pixel Interdependency'', First International Conference on Electronics, Circuits, and Systems (ICECS'1994), pp. 58-63, December 1994, Cairo, Egypt.

Abstract

Given a row-wise (column-wise) digital image of unknown width (height) as a one-dimensional sequence of pixel values, it is required to find its width (height) in order to construct it as a two-dimensional image. This paper proposes a solution for this format-detection problem using the entropy per pixel as a measure of interpixels correlation. The proposed solution is applied successfully on different images. The results show that the entropy measure successfully finds the image format. Since the proposed method relies on the correlation between pixels, it is still successful in the case of blurred or high pass filtered images or in the presence of noise.