Google+ Facebook LinkedIn ResearchGate DBLP Google Scholar

Research Interests

My current research interests are (but not limited to):

  • Large-Scale Content-Based Image Retrieval
  • Data Mining (text and images)
  • Biomedical Informatics
  • Machine Learning
  • Data Science
  • Big Data Analysis
  • Image Processing
  • Sola Image Data Analysis
  • Classification Algorithms

In my current position, I use machine learning, data mining, and medical ontologies to discover hidden prescription drug interactions and build predictive models to enable the learning health system. My duties involve the transformation and automated annotation of the Stanford STRIDE electronic health data records (structured and unstructured data) into research databases that could be used by other researchers in our lab for their work. In this position, I have successfully acquired invaluable domain knowledge in biomedical informatics and have collaborated with other labs in our department, as well as other international organizations, building an extensive research network in the field and developing community-wide tools for electronic phenotyping.

I have spent my past years analyzing Solar Image data from the TRACE and SDO missions, and while doing my Ph.D. dissertation we developed in conjunction with the SDO CBIR system (under deployment) a tool named imageFARMER to build custom made Content-Based Image-Retrieval (CBIR) systems for almost any image domain/dataset. This tool is available here. During this process I have worked with not only Solar, but also medical and natural scene images. While our software tool is designed to work specifically for Solar images, it has show surprising results for Medical datasets as well (read our ICIP 2011 publication here).

If you have any image datasets that you want to see if a CBIR systems will serve your retrieval needs, do not hesitate to contact me (here). We are always looking for new and bigger datasets. Our current SDO dataset grows by 70,000 images per day, making it one of the current largest image datasets available.

I am also open to any collaborations (academic or industry) for the topics outlined at the beginning of this page or any other interesting Data Mining / Image Processing / Machine Learning problems that you can come up with. I consider my-self a very approachable guy, so just shoot me an email and let's start talking.