Single Device-Based Deep Learning Approach for Geographical Position Spoofing Detection on an Instant Messaging Platform

With the quick growth of digital technology, image transferring is the most common process over the internet. Now, location registration on mobiles has become trendier in social media. At the same time, the criminals can modify the geo-position location as per their needs. Hence, there is a need to detect the genuineness of geo-position. Currently available techniques for spoofing detection fail to detect prior information and they are effective for small databases. Hence, the efficient deep learning-based geo-position spoofing detection technique is developed for the IM platform. The residual noise that exists in an image is mined from the input image by a DRN, which is trained by developed HBSMO. Then, the camera footprints are extracted from the noise-free images by the fuzzy filter. Based on the extracted camera footprints, the spoofed image is detected using the RV coefficient, and finally, DCT and correlation are utilized to determine which image is the spoofed image. Finally, the developed HBSMO-based DRN approach is assessed for its effectiveness using several measures, such as testing accuracy, FPR, and TPR, and is obtained that the proposed method has attained values of 0.92, 0.89, and 0.91, correspondingly.