|Title||Estimation of average cell shape from digital images of cellular surfaces|
|Publication Type||Conference Paper|
|Year of Publication||2004|
|Authors||Iles, P., D. A. Clausi, and G. W. Brodland|
|Conference Name||1st Canadian Conference on Computer and Robot Vision|
|Conference Location||London, Ontario, Canada|
|Keywords||adaptive contrast enhancement, composite cell, shape detection, watershed segmentation|
A method is proposed to determine the average cellular geometry in high-resolution images of embryonic epithelia. The concept of a ‘composite cell’ is used to represent the average cell shape. This composite cell can provide evidence of stresses present in the epithelia. A new adaptive contrast enhancement routine is applied to the input image first, followed by an iterative watershed segmentation. The composite cell is calculated from the segmentation results. Qualitative results show that this computationally inexpensive algorithm produces accurate results for a variety of image sizes, contrast levels, cell shapes and appearances.