DETECTING TAMPERED IMAGE BASED ON CONTRAST ENHANCEMENT OF Y CHANNEL

During the image tampering, contrast enhancement operation is often used to highlight some information, weaken or remove some unwanted information in images. However, this operation simultaneously leaves specific fingerprints in the image's pixel value histogram. This paper presents a wavelet analysis method to detect tampering image. We firstly convert the RUB color space into YCbCr color space, and extract the Y monochrome channel image; secondly, the normalized energy is calculated in the wavelet details sub-bands after wavelet transform of image's pixel value histogram of the component; finally, the altered image will be identified according to the normalized energy. The contrast experiment results about wavelet transform and Fourier transform indicate that the former is better than the latter in both false positive rates and time complexity. In addition, the proposed wavelet analysis approach can be applied to detect local contrast enhanced image, too. The results of the experiment show there is an obvious distinction between tampered and unaltered sub-blocks divided.