vbSPT, an acronym for variational Bayes single particle tracking, is a software package for analysis of single particle diffusion trajectories, where the diffusion constants switch randomly according to a Markov process. The program runs on Matlab, but uses compiled C-code to speed up the most computer intensive loops.
The main advantage of vbSPT is the ability to learn the number of diffusive states directly from the data, in addition to model parameters such as transition rates and diffusion constants. The statistical basis for this feature is an approximate Bayesian approach to hidden Markov models known as variational Bayes .
If you use vbSPT in your research, please cite our work .