Facial Image Denoising Using Convolutional Autoencoder Network

This project implements and reproduces the key findings from the research paper:
"Facial Image Denoising Using Convolutional Autoencoder Network"
by N. M. Tun, A. I. Gavrilov, and N. L. Tun (IEEE Xplore Link).

---

📖 Overview

Noise in facial images significantly impacts the performance of face recognition systems, especially in outdoor or uncontrolled environments. This project presents a deep learning-based denoising method using Convolutional Autoencoders (CAEs) to improve image quality prior to recognition. The solution is trained and evaluated on the ORL face dataset, serving as a robust preprocessing module for facial recognition pipelines.

---

🧠 Based On

Citation:
N. M. Tun, A. I. Gavrilov and N. L. Tun, "Facial Image Denoising Using Convolutional Autoencoder Network," 2020 International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM), Sochi, Russia, 2020, pp. 1-5.
DOI: 10.1109/ICIEAM48468.2020.9112080

---

🛠️ Tech Stack

• Python
• TensorFlow / Keras
• NumPy, Matplotlib
• ORL Face Dataset

---

💡 Key Contributions

• Implements a convolutional autoencoder architecture for facial image denoising
• Trained and validated on the ORL face database
• Demonstrates significant improvement in image clarity under noisy conditions
• Designed as a preprocessing stage for face recognition systems