Forgery-Detection
The desire for authenticity and to protect integrity drives the identification of forged images and signatures. Therefore, we have built an application that can detect image and signature forgery.
Image forgery refers to the alteration of a digital image (copy-move image forgery detection) to hide some important information while signature forging is the illegal act of imitating another person's name or signature on documents.
These can be effectively overcome using our project.
Each person may have a distinctive signature. However, because two signatures made by the same individual could resemble one another quite a bit, signatures pose a variety of difficulties. Even across signatures made by the same person, many aspects of a signature can vary. Again, image forgery detection is a technique for locating and identifying fabricated elements in a modified image. To determine whether a given image is manipulated or altered in the original, a sufficient number of attributes must be present to categorize it as either forged or non-forged.
In image forgery detection copy-move forgery detection is performed. In this type of forgery, an attacker copies a region of an image and pastes it into another region, often with the aim of hiding or adding information. This technique is commonly used in tampering with images, such as in instances of manipulating evidence in forensic investigations, creating fake documents, or even in photo editing software. The solution performed is by applying a DCT transformation, lexicographic ordering, Euclidean operations, and finally eliminating pixels to produce a mask.
Signature forgery detection is the process of identifying and verifying the authenticity of a signature on a document or other written material. This type of forgery occurs when someone attempts to replicate or forge another person's signature, often with the intention of fraudulently using the signature to gain access to resources or commit fraud. The solution used to detect a forged signature is MLP (Multi-layer Perceptron). The learning process of MLP involves adjusting the weights of the connections between the neurons in order to minimize the error between the actual output and the desired output.