Image-Denoising

With CCTV images not being very clear on zooming there is a great demand for image denoising models. Build a model which takes input of blurred RGB images and outputs clear images. Carefully study the kind of noise CCTV images have and target accordingly.

We implemented ADNet by customizing it and trained them on Google Colab.

Dataset

Div2k

Technology Used

• PyTorch (Implemented ADNet)
• Sci-Kit Image (To Add Random Gaussian Noise to the images)

References

• Methods for image denoising using convolutional neural network: a review
• ADNet
• PyTorch Tutorial
• Attention-guided CNN for image denoising