北京北京大学人工智能AI6.ppt

Forensics MotivationSeeing is believing … or is it? Easy to be deceived Goals Identify image tampering methods. Assess methods available for protecting images. Assess image authentication techniques. Identify directions for future work. Categories of Image Tampering There are three main categories of image tampering: Enhancing Compositing Copy/Move Enhancing Changing the color of objects Changing the weather conditions Blurring out objects Compositing Copy-Move Copying regions of the original image and pasting into other areas. The yellow area has been copied and moved to conceal the truck. What can be done to protect digital images? Watermarking Fragile watermarks Semi-fragile watermarks Self-embedding watermarks Digital cameras with watermarking capabilities Digital Fingerprinting/Signatures Digital cameras with fingerprinting capabilities Digital Watermarking The basic concept of digital watermarking an image is that a low level signal is placed directly into the image data. Any manipulation of the image will impact the watermark and subsequent retrieval of the watermark and examination of its condition will indicate if tampering has occurred. Fragile Watermarks Fragile watermarks are designed to detect every possible change in pixel values . Variety of Techniques but in most cases, the watermark is embedded in the least significant bit (LSB) of the image. Advantages: Pick up all image manipulations – malicious and non-malicious Disadvantages: Too sensitive Passive Authentication Techniques Detecting Copy-Move Detecting Traces of Re-sampling Detecting Light Inconsistencies Copy-Move Detection Copy-Move Detection Re-sampling Detection Inconsistencies in Lighting Conclusions Digital image forgeries can be used to deceive the public and the authorities. They are here to stay. Until non destructible/ non removal digital watermarks are perfected, passive authentication will remain necessary. Currently no single passive authentication technique can d