Information retrieval from images to tackle health and security problems

Come to learn, network, and socialize. Food will be provided by Business Intelligence Solutions (http://busintellsol.com/)! Here's an outline of the agenda:

7:00-7:30 pm: Networking session, drinks and pizzas are available to consume

7:30-8:05 pm: "Extracting rich information from biological images to tackle world health problems" by Anne Carpenter (Director, Imaging Platform, Broad Institute)

8:05-8:40 pm: "DICOS: The case for standardized data in security" by Lorena Kreda (Machine Vision Consultant, specializing in aviation security)

______________________________________

Extracting rich information from biological images to tackle world health problems

by Anne Carpenter

Director, Imaging Platform

Broad Institute

Abstract: Microscopy images contain rich information about the state of cells and organisms and are an important part of experiments to address a multitude of basic biological questions and world health problems. Our laboratory works with dozens of collaborators around the world to design and execute image-based experiments, primarily high-throughput screens. These experiments test thousands of chemical or genetic perturbations in order to identify the causes and potential cures of disease. High-throughput experiments are becoming a major source of insight in the quest to systematically identify novel drugs and regulators of important biological processes.

Machine-learning approaches, guided by a biologist’s intuition, have been particularly successful for measuring subtle aspects of cells’ appearance. We are also taking a systems biology approach to explore the potential of extracting patterns of morphological perturbations (“signatures”) from cell images in order to identify the similarities between various chemical or genetic treatments. The goal of these experiments is to identify distinctions between human isoforms of cancer-relevant proteins, mechanisms of liver toxicity, and diagnostics for bipolar disorder and schizophrenia. Our long-term goal is to make perturbations in cell morphology as computable as other sources of large-scale functional genomics data.

The algorithms and approaches we develop are freely available through the biologist-friendly open-source software, CellProfiler, for both small- and large-scale experiments.

Biography: Dr. Anne Carpenter directs the Imaging Platform at the Broad Institute of Harvard and MIT, a non-profit biomedical research institute. Dr. Carpenter leads a team of computer scientists and biologists to develop advanced methods to quantify and mine the rich information present in cellular images. Typical projects involve testing hundreds of thousands of samples by microscopy, helping to determine the functions of genes and to identify chemicals for use in research and as potential therapeutics. The group maintains the open-source software project, CellProfiler, which is used worldwide by the scientific community and was awarded the Bio-IT World Best Practices Award in 2009.

In 2008, she was elected the youngest fellow of the Massachusetts Academy of Sciences in recognition of her work. She has been awarded research grants from the US National Institutes of Health, the National Science Foundation (CAREER award), the Howard Hughes Medical Institute, and the Life Sciences Research Foundation. She has been named a “Rising young investigator” by Genome Technology magazine and was featured in a public television special, “Bold Visions: Women in Science & Technology”.

After earning her B.S. from Purdue University in 1997 and her Ph.D. from the University of Illinois at Urbana-Champaign in 2003, Dr. Carpenter completed postdoctoral work with David Sabatini at the Whitehead Institute for Biomedical Research, co-mentored by Polina Golland at the Computer Science/Artificial Intelligence Laboratory at the Massachusetts Institute of Technology.

DICOS: The case for standardized data in security

by Lorena Kreda

Machine Vision Consultant, specializing in aviation security

Abstract: It is undisputed that openness and information sharing helps foster innovation, and the evolution of medical imaging solutions since the adoption of the DICOM (Digital Imaging and Communication in Medicine) data standard in the early 1990s is one example that supports this claim. Image-based explosives detection systems have emerged over the past 2 decades, largely based on X-Ray technology borrowed from medical imaging, yet currently relies on widely varying proprietary data formats. The security industry is just starting to adopt its own data standard based on DICOM called Digital Imaging and Communications in Security (DICOS), and this will help create opportunities for entities other than just the equipment manufacturers to participate in innovative explosive detection solutions and enhancements. This talk will introduce the standard (briefly) and explain the benefits of adoption in the context of image processing challenges for explosives detection systems.

Biography: Lorena Kreda has been employed in the aviation security industry for 12 years, having worked at Vivid Technologies/PerkinElmer Detection Systems and Reveal Imaging Technologies, prior to becoming an independent consultant in April 2012. Ms. Kreda provides algorithm and image processing and system engineering expertise to clients primarily in the aviation security industry. From 12/2009 to 3/2012 Ms. Kreda served as the Director of the Machine Vision Group at Reveal, an SAIC Company. In this role, she led multiple explosive detection algorithm development efforts on Reveal’s multi-view and CT X-Ray systems and several key EDS certifications including HMEs and liquid explosives in the US and Europe. Ms. Kreda holds Bachelors and Masters degrees in Electrical Engineering from Worcester Polytechnic Institute and completed 2 years of Ph. D. coursework and research in Electrical Engineering at Northeastern University.