Cell Parameter Calculation using Computer Vision



We are inclined to measure the physical characteristics of cancer cells to monitor their growth rate/ survival time. We did an experiment with HeLa cells that can be seen to the right.

In order to detect the cells we use a network flow algorithm (blob detection). However for the algorithm to be able to work more effectively, we manipulate the image so the algorithm can detect discrepancies easier.  We used the following image processing techniques and filters:Image processed HeLa

  • Image smoothing
  • color thresholding
  • mDetectionorphology
  • addition to standard deviation of the image

The blob detection algorithm, looks for areas of differ in color and brightness and discriminates against sudden changes of these characteristics if they happen in a region. The processed image is passed to this algorithm and we are able to calculate the area and perimeter of  each cell and compare it to the previous measurements to find the rate of growth.