Design and implementation of light DNN architecture for organ segmentation on abdominal CT scans Design and implementation of light DNN architecture for organ segmentation on abdominal CT scans

The objective of the thesis is to investigate the implementation and design of light DNN architectures that can be easily integrated on memory constrained hardware without compromising the accuracy. The usage of neural architecture search (NAS) approach is desired.

This project will be done in collaboration with the technological start-up Sycai Technologies