Detection And Implementation On FPGA Of Forged Digital Images Based On An Improved Algorithm

Today,digital image,as a common multimedia media,plays a increasingly vital role in modern society.However,with the development of image editing technology,image forgery has become easier,and even ordinary people can complete image forgery.When these forged images are used in important occasions such as court trials,etc.,it will cause a series of serious problems.Therefore,it is important to detect whether an image is forged.Passive digital image forensics technology does not require prior knowledge,and image detection is based only on the content of the image itself.Image copying and pasting tampering is a very common form in many image tampering methods.Common methods are difficult to distinguish and have become a hot topic of research.This paper studies a classic improved image copy-paste tamper detection algorithm.First,an improved algorithm for this topic is given,by reducing the contrast threshold to obtain more feature points.Then,using Haar wavelet instead of the gradient histogram information of the feature points to generate the feature vector,the dimension of the feature vector is reduced to 64,which is half of the classical SIFT.This can effectively reduce the time complexity of feature vector generation and subsequent feature matching.Then,using the g2 NN matching algorithm for feature matching,it is possible to detect image copy and paste tampering of multiple targets.Finally,based on a clustering,the GMS algorithm is used to perform quadratic clustering to further improve the accuracy of algorithm detection.This paper uses the ML605 board to complete the FPGA implementation of the improved detection algorithm.First,use the Windirver software to generate the INF file.After the installation,the Windirver software is used to generate the basic PCIe driver framework for fast programming of the PCIe driver.The PCIe driver is used to transmit the image data to be detected from the host computer to the ML605 board through the PCIe interface,and then transmit the detection result back to the host computer through the PCIe interface,and finally display the detection result with Matlab.The FPGA implementation of the improved detection algorithm includes feature point extraction,feature vector extraction and feature matching.Finally,the FPGA implementation and Matlab simulation results of the improved detection algorithm are compared,and the results are analyzed and illustrated.