Image Tamper Detection using Mathematical Morphology

This paper proposes a tamper detecting method for images using mathematical morphology. The proposed method utilizes the idempotent property of morphological operations rather than fragile watermarking methods. While fragile watermarking methods that must extract the embedded watermark and compare it with the possible watermark to detect tampers, the proposed method detects and localizes tampers by only morphological operations and image subtraction in the spatial domain. Moreover, a visual signature system is proposed based on the proposed tamper detecting method. Experimental results show the effectiveness of the proposed method.