Allard Hendriksen, researcher at <u>CWI’s Computational Imaging group</u> was awarded the Best Poster Prize at the “Mathematics of Machine Learning” symposium organized by the London Mathematical Society and the University of Bath. He won the prize for his poster titled Noise2Inverse: Deep tomographic denoising without high-quality target data.
In X-ray computed tomography, a 3D model of the interior of an object is computed from a sequence of X-ray images. As the exposure time of the X-ray acquisition is reduced, noise is introduced in the images and the reconstructed model. For removing this noise, deep convolutional neural networks have been shown to be effective, but have so far required a dataset of noise-free target images for training. Our research suggests that it is possible to train such networks without any additional noise-free data by changing the training strategy. This opens the doors for application of deep convolutional neural networks in applications where obtaining noise-free images is infeasible, such as battery research and tomography of quickly evolving dynamic systems.
Mathematics of Machine Learning Symposium
The conference advocates the connection of many mathematical disciplines like numerical analysis, inverse problems, optimisation, statistics, optimal transport, dynamical systems, partial differential equations to ML. By bringing together world-leading mathematicians, statisticians and data scientists to discuss recent developments in the fundamental understanding of ML it aims to shed light into the mysterious pathways of ML.