Image retouching can have varying degrees of impact on people‘s physical and mental health. When calling for reduced image retouching, the media should be supervised to minimize image retouching or to mark retouching images. In the supervision, image forensics methods are indispensable. For the local deformation of the image, this paper finds that the method based on BAG can detect the image tampering. The author adds the information of different color channels based on BAG to better locate the deformation. At the same time, this paper puts forward an idea about the detection of deformation direction based on block effect.