Comparing analysis engines for generated micro-patterned, actin images

Actin fibers are vital in cell development, growth and interactions. Recent advances in micro-patterns force cells to conform to specific shapes which is a huge benefit to image analysis. In this work, cell images conforming to micro-patterns are computer generated through the minimization of four energy terms, which in turn may provide evidence of the mechanics at work in the biological cells. Two methods of comparing original images to the generated images are considered and their performances compared. The conclusion is that a superior performance is obtained with a method based on the mammalian visual cortex over that of traditional Gabor filter analysis. Further analysis leads to an understanding the forces that generate optimal images which can be extrapolated back to the original biological sample in support of understanding actin behavior under the influence of drugs.