Mohammad Emad

Eindhoven University of Technology
Electronic Systems Group

EDL P16-25 P4: Deep Learning for High-Tech Systems and Materials

Research assignment:
Efficient deep learning for image restoration

This project concerns the application of DL for high-performance electron microscopy, in close cooperation with Thermo Fisher company. Electron microscopy enables e.g. the study of biological tissue properties and cell structures, to obtain in-depth knowledge on disease properties and possible treatments. This project focuses on using deep learning for 3D image restoration tasks like super-resolution, denoising, inpainting and segmentation of electron microscopy images.
Electron microscopy images are usually very big and applying a deep neural network to the whole image consumes a lot of time and energy. Efficient deep learning platforms can be highly beneficial for this use case. Lack of training data is another challenge for image restoration of electron microscopy images. It forces us to use unsupervised or semi-supervised techniques rather than supervised methods.

The network of our proposed super-resolution method (DualSR)


DualSR: Zero-Shot Dual Learning for Real-World Super-Resolution.
Mohammad Emad, Maurice Peemen, Henk Corporaal. Conference proceedings of Winter Conference on Applications of Computer Vision. Open access-green.

Personal information