Fake Image Detection on Social Media using CNN Algorithm

DOI

No DOI

Published In

Volume 4, Issue 3 (2023)

Published Online

May-June 2023

Pages

01-05

Article Type

Research Article

Check for updates

No DOI available for cross-reference

Abstract: In today's time, the common man is being fooled through fake pictures because they do not know whether the pictures are real or not.In this technological age, people have placed social media at a prominent level in their daily lives. Most of the people share their information or any important thing on social media through text message image and video like twitter, snapchat, Face book, WhatsApp telegram and many more. Biometric techniques are helpful now for identifying people, but criminals alter their look, behaviour, and psychology to trick the identification system. In order to solve this issue, we are employing a novel method called Deep Texture Features Extraction from Images, followed by the construction of a machine learning model utilising the CNN (Convolution Neural Networks) algorithm. As it largely relies on features extraction utilising the LBP (Local Binary Pattern) method, this technique is often referred to as LBP Net or NLBP Net. The sole purpose of this research is to create a model that can be used to classify social media content to detect any threats and fake images. This model was made using Deep Learning which is Convolutional Neural Network (CNN). LBPNET, a machine learning convolution neural network, is the name of the network we created for this research toidentify fraudulent face photographs. Here, we will first extract LBP from the photos, and then we will train the convolution neural network on the LBP descriptor images to produce the training model. Every time a new test picture is uploaded, the training model will use that image to determine if the test image contains fraudulent images or not. Details regarding LBP are shown below. Now that the feature vector has been processed, it may be classified using a machine learning method such as the Support Vector Machine, extreme learning machines, or another one. The study of textures or the recognition of faces may both be done using these classifiers. The results of this research will be helpful in monitoring and trucking of images in social media to detect unusual material counterfeit images and protect social media from threats and fraudsters.