Toward Reliable Image Forensics
Choudhary Shyam Prakash, Sahani Pooja Jaiprakash, Naween Kumar · 2024
Social media platforms are crucial for broadcasting news globally, but they also contribute to the spread of fake news, which defames individuals or organizations. Images are the most common medium for information transmission, and sophisticated tools can manipulate images effectively making it impossible to detect manipulation with our eyes. This chapter outlines some techniques for determining the authenticity of an image. Active and passive methods are the categories into which image forensic techniques are divided. In recent years, academics have developed models for image authentication using machine learning and deep learning techniques. Traditional methods struggle with large-scale data classification, while deep learning algorithms offer improved performance but require careful hyperparameter selection. Researchers have proposed proprietary architecture and transfer learning models for better precision in image forensics. Experimental results from previous publications demonstrate the applicability of deep learning models in image forensics.