This repository contains codes used and relative documentation for the 2nd project of Imperial College London MRes Biomedical Research (Data Science stream) in 2019-2020 academic year. Project thesis can be accessed here.
-- Conducted by: Mingze Gao.
-- Supervisors: Dr. Joram Matthias Posma & Dr. Katia De Filippo
The overarching aim of this project is to implement ML/DL based algorithms (fastER / Mask-RCNN) to,
- Analyse splenic IVM video data
- Compare cell segmentation performance with Imaris
- Use statistical metrics to quantify the migration characteristics of neutrophils
- Investigate Neutrophil-B cell interaction behaviours under different experimental conditions.
All codes are implemented and tested in Python 3.7.3 and Conda 4.8.3
Programs are ran locally on Windows 10, Intel(R) Core(TM) i7-8565U CPU @ 1.80GHz, RAM 8.00GB.
Deep learning codes are connected to HPC based on NVIDIA Tesla K80 GPU with 24GB memory.
Multiple object tracking is based on SORT, codes are revised and acknowledged approriately inside thesis to fulfill project needs.
Uploaded codes contains all the scripts in this project ranging from, data pre-processing, extraction, cleaning, cell trajectory tracking, migration and interaction analysis, to data visualisation (raw data excluded due to the nature of the experiment).