3DeepLearning
Title: 3DeepLearning
SNIC Project: Berzelius-2021-44
Project Type: LiU Berzelius
Principal Investigator: Per Uhlén <per.uhlen@ki.se>
Affiliation: Karolinska Institutet
Duration: 2021-09-17 – 2022-04-01
Classification: 30203
Homepage: https://github.com/uhlen-lab
Keywords:

Abstract

This project is a continuation of our recent publications that show that 3D microscopic imaging can more accurately diagnose tumor samples than current 2D methods used in the clinic (Nature Biomed Eng 2017, Nature Biomed Eng 2020, Trends in Cancer 2018). Pathologists have classically diagnosed cancers using 2D microscopy and staining with dyes or antibodies to study specific structures, molecules, and/or proteins. However, studying 3D structures such as tumors in 2D creates an information gap between the recorded data and the original tissue. Recently, new protocols that render biological tissues transparent or clear to visual light have been developed. This has promoted the development of new imaging techniques, e.g., light-sheet microscopy, and advanced image processing algorithms that we will use in this project. In this project we will use the advances in 3D imaging and develop new deep learning image processing tools.