Microscopic Image Analytics: A Model-based learning Approach

abstract

Microscopic image analytics is becoming increasingly important due to the current explosion data, the many biological applications, the role in drug discovery and evaluation, and the ability to quantitate and discover complex spatiotemporal cell properties at multiple scales. In this talk I will present a range of computational methods based on coupled modeling and learning methods to address complex problems in microscopic image analysis. I will first present methods for cell tracking, classification, and cell shape analysis for cancer stage classification. I will then present fast learning methods based on hashing for histopathology analytics and will conclude with aggregate cell tracking based on bioluminescence. Finally, I will conclude with future directions in this increasingly important area.