@conference {358, title = {W01-03: Computer Vision for Next Generation Phenomics and Tree of Life}, booktitle = {International Conference on Biomedical Ontology and BioCreative (ICBO BioCreative 2016)}, year = {2016}, month = {11/30/16}, publisher = {CEUR-ws.org Volume 1747}, organization = {CEUR-ws.org Volume 1747}, address = {Corvallis, Oregon, USA}, abstract = {

To build the Tree of Life, scientists collect data on all heritable features {\textendash} both genotypes (e.g., DNA sequences) and phenotypes (e.g., anatomy, behavior, physiology) for all living and extinct species.\  The collection of phenomic data for tree-building has lagged far behind the collection of genomic data. Advances in computer vision have the potential to change this situation. In this talk, I will present a computer vision system developed in our lab for extending phenomic matrices, and in this way\ building the Tree of Life.\ Rows of a phenomic matrix represent images of specimens belonging to various species of interest, and columns represent scores of their phenomic characters. Given a\ phenomic matrix where only few rows are manually annotated with character scores, our vision system extends the matrix row-wise by populating missing character scores of the remaining species in the matrix. The talk will present our experimental results on scoring phenomic characters in images of bat skulls, nematocysts, and leaves, available in the Morphobank and Bisque data repositories.

}, keywords = {Bisque data repository, data visualization, Morphobank, Phenotypes, Tree of Life}, url = {http://icbo.cgrb.oregonstate.edu/}, author = {Sinisa Todorovic} }