Implementing bayesian networks in python for gaze estimation using visual saliency I am developing an appearance based gaze estimation system based on opencv and python.
I have currently developed a prototype which can estimate the gaze based on active calibration, which is cumbersome at times.
The following paper presents a technique based on visual saliency map for implicit training of the system based on bayesian networks-http://www.ecse.rpi.edu/~cvrl/chenj/1669.pdf
I have already obtained saliency map for a couple of images from MIT archives
but am unaware on how to implement it in the algorithm mentioned as it involves indefinite integrals.
And as far as Bayesian networks are concerned I am a noob.
I already tried looking for similar implementations but could not find any, will anyone be kind enough to point me to some material to understand and start implementing the system myself?
I am pretty new to machine learning, so forgive me  if the question is trivial
 A: You may wish to have a look at the original model of Itti & Koch (2000). The details are in Itti, Koch & Niebur (1998). Koch's Matlab implementation is available from http://www.saliencytoolbox.net. Itti's Lab has a C++ implementation which targets performance: http://ilab.usc.edu/toolkit/.
The original work does not use graphical models and you will need to consult Harel, Koch & Perona (2006) on the details of the graph-based implementation. A brief look at Jonathan Harel's profile page shows that he has a Matlab implementation of the graph-based approach posted online: http://www.vision.caltech.edu/~harel/share/gbvs.php
Literature
Harel, J., Koch, C., & Perona, P. (2006). Graph-based visual saliency. In Advances in neural information processing systems (pp. 545-552).
Itti, L., Koch, C., & Niebur, E. (1998). A model of saliency-based visual attention for rapid scene analysis. IEEE Transactions on Pattern Analysis & Machine Intelligence, (11), 1254-1259.
Itti, L., & Koch, C. (2000). A saliency-based search mechanism for overt and covert shifts of visual attention. Vision research, 40(10), 1489-1506.
