Fluorescence Image Enhancement for Digital Pathology (DP) Scanners

Fluorescence imaging is a powerful technique to focus and acquire high-resolution pathology images that contain rich sources of information that is useful for diagnostics in clinical trials to study possible disease such as cancer glands and tumors. Despite the high definition images, the collected images are severely contaminated with noise artifacts which make most feature detection algorithms susceptible to such degradations. Besides, due to over-sized problem, a computational efficient algorithm is needed to process such data. The main objective of this proposal is to introduce an image analysis software that is capable of addressing two main challenges: first is to introduce a denoising algorithm to overcome both issues of low signal-to-noise-ratio (SNR) and super-resolution problem. The second goal here is to localized tissues in fluorescence images for segmentation. This makes the storage of the data much more efficient in compressed format for any retrieval image processing tasks.

Faculty Supervisor:

Konstantinos (Kostas) Plataniotis

Student:

Mahdi Hosseini

Partner:

Huron Digital Pathology

Discipline:

Engineering - computer / electrical

Sector:

Medical devices

University:

University of Toronto