# CS 585 Project: Tracking of Living Cells in Microscope Images

Boston University
Instructor: Prof. Margrit Betke
Departments of Computer Science and Mathematics
Yizhang Bai
Teammates: Haoyu Zhang, Hui Li
11/7/2016

## Project Proposal (General)

Our goal is to distinguish the condition of a cell and determine whether this cell is healthy or not.

Since every cell has its own shape and own moving rule. Thus when we track a cell and observe its moving trail and shape, we can determine whether this cell is normal or not. At the same time, when tracking the living cells, we can observe it and determine its age, health condition, etc.

## Project Proposal (Detailed)

• Project problem definition:

Implement a system that would follow the movements of a living cell in time-lapse phase-contrast microscopy video. A first task would be to track the position of the centroid of the cell. Another task would be to analyze the changes of the shape of the cell, following the analysis discussed in class that uses axes of first intertia. An exciting extra step would be to register the outline of the cell from image to image, using deformable models or non-rigid registration techniques discussed in class.

1) We need to find/design an algorithm to track the position of the centroid of the cell.

Current formula for Skeleton:

(S* is the track for points that will remain as the "skeletion", s is a point, s_bar is the center of the mask window, d(s,s_bar) is the distance between s and s_bar, N is the mask window)

2) We also need to find/design an algorithm to analyze the changes of the shape of the cell, following the analysis discussed in class that uses axes of first intertia..

Probable methods:

Kalman filter, Hungarian algorithm

3) Also, if possible, we need to register the outline of the cell from image to image, using deformable models or non-rigid registration techniques discussed in class.

• Images or link videos that we will use as our input:

C2DL-MSC

C3DL-MDA231

N2DH-GOWT1

SNPHEp2_images

(Example cited from datasets. )

• Inputs and Desired results:

• What we've done so far:

Currently, we are gathering information that might be helpful in this project. (including reading papers and notes in "Potential Related References")

Also, we are trying to implement some of the technologies needed for this project.

• ## Potential Related References (Informal Citing):

CS 585 Project Guidelines for Fall 2016

CS 585 class notes, including:

ROC Analysis

Segmentation

Thresholding

Multiple Hypothesis Tracking

Optical Flow

Absolute Orientation