CNN-LSTM Model for Deepfake Image Detection
Reena S. Satpute, Chidozie Peter Onwe
Abstract
Deepfakes are a powerful tool for producing artificial media that can convincingly mimic real content, raising concerns about its potential misuse for spreading misinformation and manipulating public opinion. Detecting deepfake images has become a critical research area, with numerous techniques developed using deep learning methods. This study presents a strategy for spotting deepfake images with a model that combines convolutional neural networks (CNNs) and long short-term memory (LSTM). The combination blends CNNs' spatial awareness with LSTMs' grasp of temporal context. Through successful performance on accessible datasets, the developed approach achieves an accuracy rate of 98.20%, showcasing its ability to accurately discern deepfake images while maintaining a low false-positive output. With an error of 0.28%, the model highlights the complexities and challenges in deepfake detection. The results emphasize the effectiveness of utilizing combined deep learning techniques in addressing the critical problem of precisely detecting manipulated images.