Presented in London
The paper “Learnable Total Variation with Lambda Mapping for Low-Dose CT Denoising” was presented during IEEE ISBI 2026 in London.
For the project team, this was more than a conference appearance. It was also a first major academic milestone for Yusuf Talha Basak, whose work on interpretable low-dose CT denoising reached one of the most established venues in biomedical imaging.
What the paper is about
The paper connects a classical model-based prior with learned spatial adaptivity. Instead of abandoning physical interpretability entirely, the method keeps total variation regularization in the loop and learns a lambda map that changes the regularization strength across the image.
That gives the method a useful middle ground: more transparent than a generic black-box denoiser, but more adaptive than a fixed handcrafted prior.
Acknowledgment
Special thanks go to Nureddin Kamadan, who presented the work in London and represented the project onsite with care.
Event details
- Conference: IEEE International Symposium on Biomedical Imaging (ISBI) 2026
- Dates: 8-11 April 2026
- Location: London, UK
Repository and additional public material can be linked from the site once they are ready for release.