An Advanced Digital Image Forgery Detection via Local Binary Pattern and Haralick Features

Nowadays, authorizing the genuineness of an image is a big challenge. For ensuring the trustworthiness of images, a secure hashing method is developed. This method detects the tampering like addition, removal and replacement of any object from the image, unusual color modifications, resizing, and image compression and especially locates the forged area. A hash sequence is generated by extracting features from reference and test image. Global features are extracted using Zernike moments representing luminance and chrominance characteristics. Texture features are extracted by dividing the image into non-overlapping blocks. Haralick features and Modified Local Binary Patterns are extracted from each block. Secret keys are introduced for hash construction. The hash is sensitive to malicious forgeries. By decomposing the hashes the type and location of forgery can be determined. The proposed method will combat the problem of image forgeries in various domains like legal services, medical images, forensics, intelligence and sports.