A rational approach to noise discrimination in video microscopy particle tracking

The authors describe a rational approach to an important step in video microscopy particle tracking called noise discrimination. Using the morphology of the brightness profiles produced by the particles, false and dubious detections are eliminated from the valid tracking data. This selection process is found to affect the spatial resolution epsilon and the depth z(b) of the tracking. Accordingly, it is shown that a stringent selection of the particles closest to the focal plane improves the spatial resolution (i.e., decreases epsilon) but also diminishes z(b). The authors further demonstrate that the statistical accuracy is worsened by lower z(b). Noise discrimination thus has a critical influence on the overall performance of the particle tracking technique. (C) 2008 American Institute of Physics.