Topic 2: Non-Local means for image denoising¶

1. Read and discuss the paper by Buades et al..
2. Try different noise models (e.g. Gaussian, salt-and-pepper and shot noise) on different test images (acquire them yourself) with different strength and try to denoise them.
3. In addition to NL-means, implement two other denoising methods of your choice and compare the results.
4. Investigate which denoising method is most suitable depending on the applied noise.
5. Quantitatively assess the performance of your denoising implementation, i.e. visualization of the denoised images is not enough for comparison.
6. In addition to standard test images, try to acquire some noisy images using your own camera and then denoise them.

References

A. Buades, B. Coll, J.-M. Morel: A non-local algorithm for image denoising. Proc. 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2005, San Diego, CA, June 2005), pp. 60–65.