About

This project is a senior capstone project developed at CSU Bakersfield for the Senior Design Expo. This system detects image tampering in urban infrastructure photographs, where ground-level images are frequently used to monitor potential projects and development progress. We start by analyzing shadow inconsistencies using a dual-path framework that combines rule-based logic and machine learning to provide a side-by-side comparison of each method's findings. Our approach uses pixel-based algorithms to extract shadow texture, lighting strength, and depth characteristics to improve detection accuracy. We plan to develop a web application that analyzes an image's shadows and provides tampering likelihood scores with visuals of suspicious regions. Our aim is to find ways to support digital forensics and ensure the authenticity of urban imagery used in planning and reporting.