Dr Raphael Sznitman will come at MIAL on Tuesday August 26th to give a talk on his research work entitled '20 Questions to find targets in noisy biomedical image data'.
Invited talk of Dr Raphael Sznitman
Dr Raphael Sznitman will come at MIAL on Tuesday August 26th to give a talk on his research work entitled '20 Questions to find targets in noisy biomedical image data'.
Within the context of medical imaging, searching for objects of interest is a ubiquitous problem encountered at scales ranging from centimeter to nanometer. If anything, the severity of this problem has only worsened with the advent of cheaper and ever more sophisticated imaging devices, capable of producing enormous amounts of data that need to be analyzed. And while faster search methods capable of dealing with larger quantities of data are now necessary, established paradigms are showing their limits.
To this end, I will present in this talk a Bayesian formulation to the traditional "twenty questions" game, with the goal of locating objects in noisy biomedical image data. By sequentially asking a knowledgeable oracle "questions", and considering that the received answers are noisy, the goal posed is to determine a policy, or sequence of questions, that reduces the uncertainty of an object location as much as possible. We will show that principals in dynamic programming and information theory can be used to characterize an optimal policy when minimizing the expected entropy of the distribution of target locations. In particular, we will show one solution to this problem that is greedy, Bayes-optimal and simple to compute. We will then present embodiments of this concept in the context of two real applications in biomedical imaging: (i) instrument tracking during surgery, and (ii) target localization in Electron Microscopy.