An image tamper detection and recovery method based on self-embedding dual watermarking

This paper describes an image tamper detection and recovery method based on self-embedding dual watermarking. The watermark in the proposed method is a down-sampled version of an original image (cover image). Two identical watermarks are then decomposed into bit planes and embedded into the least-significant bit (LSB) plane of the cover image. This dual watermarking strategy ensures a robust performance in image tamper detection and recovery. Moreover, since the bit planes of the two identical watermarks are arranged in center-point-symmetric manner, image information is spread uniformly. This makes it possible to recover large area of tampering, such as, left, right, upper, or lower half of the cover image. Experimental results show that the proposed method is remarkably effective for detecting and recovering local image tampering, such as, image collage.