General Info
Group project for the Signal, Image, and Video (SIV) course in Artificial Intelligence Systems at the University of Trento.
- Repo: neural-prnu-extractor
- Language: Python
- Start date: 2021/11/21
- End date: 2022/02/09
Description & Objectives
This source code provides a modified version of the “FFDNet image denoising, as in Zhang, Kai, Wangmeng Zuo, and Lei Zhang. FFDNet: Toward a fast and flexible solution for CNN based image denoising.”
This version, unlike the original, concentrates on detecting the cameras PRNU.
It includes the option of training the network using the Wiener filter as a strategy to detect and extract noise from images, in addition to the original method provided in the paper.
Noise reduction is the process which consists in removing noise from a signal. Images, taken with both digital cameras and conventional film cameras, will pick up noise from a variety of sources, which can be (partially) removed for practical purposes such as computer vision. Neural-PRNU-Extractor aims at predicting the noise from an image, provided a noise level.
In addition to noise extraction, Neural-PRNU-Extractor can compute and evaluate PRNU, given a dataset of flat images, and evaluate the natural images.
Results
- Developed a neural pipeline to extract the noise from a set of camera images and perform camera identification through estimated PRNU. The performance reached was comparable with state-of-the-art technique based on Wavelet denoising.
- The neural network is based on FFDNet, which works with different noise levels and uses state-of-the-art CNN structures (PyTorch, sklearn).
Final Mark
30L (i.e. with honors) out of 30