Image Hashing Robust Against Cropping and Rotation

Image recognition is an important mechanism used in various scenarios. In the context of multimedia forensics, its most significant task to automatically detect already known child and adolescent pornography a large set images. When fighting disinformation, it identify images taken out or image montages. For this purpose, numerous methods based on robust hashing feature extraction are known, recently also supported by machine learning. However, general, these either only partially changes such as rotation pruning, they require amount data computation. We present method simple block hash that efficient compute memory efficient. To be against cropping rotation, we combine with segmentation normalize objects. Our evaluation shows produces results comparable much more complex approaches, but requires fewer resources.