Biological Shape Spaces

(object placeholder)

Cardiac dynamics segmented from 4D OCT images of the chick heart outflow tract at day 3 of incubation (cardiac cycle 370ms).

The goal of the Biological Shape Spaces project is to enable scientists to efficiently find relationships between the shape and function of biological systems, and to better understand how alterations in shape can alter function and vice-versa (see

For biological systems, shape (i.e., the morphology of an organ or organism) is frequently retrieved in the form of images, which then need to be processed and interpreted. Advances in imaging technologies are allowing an ever-increasing amount of image information with increasing resolution. Despite these advances, the analysis of images in Medicine and Biology remains largely manual, and thus time consuming and subject to operator interpretation. Quantitative, functional interpretation of shape variations remains a fundamental challenge. Computational algorithms to analyze images and shape, while available, are often ad hoc, not easily generalizable and thus usually difficult to optimize for specific applications, and frequently out of the reach of non-experts.

Advances in high-throughput imaging have led to the rapid accumulation of shape information, but the tools to analyze these data have not kept pace. The lack of a coherent framework for quantifying and analyzing biological shape has prevented the objective testing of many hypotheses that rely on morphological data. Our inability to systematically link shape to genetics, development, function, environment, and evolution has frustrated advances in biological research across multiple spatial and temporal scales, from understanding how environmental influences alter developmental morphology to interpreting adaptive responses and radiations in the paleontological record.

To breach this important gap, investigators from eight institutions, including OHSU, are collaborating in a project funded by NSF to develop the next-generation computational tools that will allow biologists and physicians to analyze and retrieve shape information automatically, and enable discovering of relationships between shape and function. The goal is to not only develop better tools for mathematically describing shape, but also methods for extracting biologically meaningful information on morphological variation. The group consists of experts in Biology, Engineering, Computer Science and Mathematics determined to develop and bring tools to the scientific community that will enable optimal and fast analysis of shape and large imaging files. The collaborating PIs in this project are: Dr. Washington Mio (lead PI) from Florida State University; Dr. Surangi Punyasena, from University of Illinois at Urbana-Champaign; Dr. Ge Yang, from Carnegie Mellon University; Dr. Rolf Mueller, from Virginia Tech; Dr. Cindy Grimm, from Oregon State University; Dr. Charless Fowlkes, from University of California at Irvine; and Dr. Sandra Rugonyi, from the OHSU Department of Biomedical Engineering.

The collaborative project was launched in November 2010, with the First Workshop on Quantification of Biological Shape, organized by Dr. Rugonyi, and hosted at OHSU. The group has developed a website (see that will facilitate search and retrieval of images and image shape data, as well as sharing developed computational tools to find meaningful relationships between biological shape and function.