# tampering_detection

**Repository Path**: HCH-AI/tampering_detection

## Basic Information

- **Project Name**: tampering_detection
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 1
- **Created**: 2025-08-29
- **Last Updated**: 2025-08-29

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Image Tampering Detection
Detecting tampered/fake images using Deep Learning

This is a rough implementation of the following forensics challenge.

Research Paper: http://openaccess.thecvf.com/content_cvpr_2018/papers/Zhou_Learning_Rich_Features_CVPR_2018_paper.pdf

Learning Rich Features for Image Manipulation Detection

Download the dataset from here: http://web.archive.org/web/20171013200331/http://ifc.recod.ic.unicamp.br/fc.website/index.py?sec=5


Phase1- Using ELA (Error Level Analysis) and standard CNN architecture using softmax to just detect whether an image is real or fake.

Phase2- Finding the exact tampered region of an image using transfer learning.(ELA+UNET,SRM+Fake Images)

Results:
Phase1- F1 score=0.88 
Phase2- Dice coefficient(custom metric) = 0.94