Skip to content

Latest commit

 

History

History
50 lines (34 loc) · 2.55 KB

README.md

File metadata and controls

50 lines (34 loc) · 2.55 KB

PRNU extraction via denoiser based on convolutional neural network (DRUNet)

Python Numpy Pillow SciPy ScikitLearn Opencv-Python Torch MIT License

Image Processing and Security

drawing

Authors

  • Mirco Ceccarelli
  • Francesco Argentieri

Relators

  • Prof. Alessandro Piva
  • Dott. Dasara Shullani
  • Dott. Daniele Baracchi

Introduction

In this project we have dealt with a branch of the Multimedia Forensics: Device Identification (know what device has taken a photo).

DeviceIdentification

In order to do this, a noise that is introduced by the sensors of each digital camera, is exploited: the PRNU (Photo Response Non Uniformity), which is different for each individual device.

Here below a brief scheme of the Device Identification process via PRNU: DeviceIdentification

Settings

Before running the code, set up the project correctly. Click here: model_zoo

In the prnu/functions.py file is used the Noise Extract method of DRUNet, if you want to use the Polimi Noise Extract method you need to comment some lines of code.

To Run

python3 example.py