Development and implementation of a custom program for post hoc strain tracking in biological specimens using...

The research presented describes the development of a custom MATLAB® program designed to examine changes in surface strain based on digital image tracking, a technique that employs markerless tracking to access regional displacement in digital images. Strain tracking is accomplished through the analysis of successive images taken during testing using a dedicated black and white camera incorporated into a tensile testing system. In the tracking program a 2-D array of grid points is mapped on the initial image from a test. Each grid point consists of a pixel array (a sub-image corresponding to a “marker”) which is then compared to consecutive images to determine the new grid point placement using cross correlation. By analyzing marker position post hoc, the number of grid points, the size of the sub-pixel array, and the spacing (longitudinal and/or regional) between grid points can be varied to account for differences in observed responses. Digital strain tracking allows for a test to be reanalyzed and eliminates the need to affix markers in specified patterns to a sample prior to testing. The new “markerless” tracking approach has been compared to the classic procedure of applying physical surface markers for tracking strain, to determine the accuracy of this technique. Preliminary testing revealed that adding texture to the specimen may be necessary (especially for low texture samples like tendons and ligaments), through the use of glitter derivatives, to increase visual contrast between multiple grid points.