Automated Cell Detection and Morphology Analysis on Microscopic Images in Imaging Flow Cytometry
Recent advances in imaging technologies have eased the collection of microscopic images, but efficient image analysis of this data remains a challenge. In the field of imaging flow cytometry, accuracy, simplicity and processing time of image processing algorithms plays a significant role. The microscopic image data acquired by video camera is of low contrast and low resolution. In addition, the camera film may be subject to varying light levels. Hence, it is crucial to develop image analysis techniques that are independent of image quality and inconsistent light levels, consequently enabling automation of their analyses. This thesis studies and implement various image analysis techniques on different subpopulations of cell that are subject to hydrodynamic pressure in microfluidic medium. It focuses on examining the credibility of these techniques by comparing against manually collected ground data. Our study targets pixel intensity-based thresholding techniques in single cell detection, cell area calculation and nucleus isolation. The studies in this thesis also indicate the advantages of implementing simpler image morphology algorithms for collecting cell morphology features used in imaging flow cytometry applications.