High-throughput scoring of image-based biological assays heavily depends on the extraction of quantitative numerical information from microscopy images. This paper describes how the Open Microscopy Environment (OME)’s MATLAB Handler combines a bioinformatics data and image repository – OME – with a numerical analysis environment – MATLAB. An OME/MATLAB coupling leverages the wealth of intellectual property invested in MATLAB algorithm implementations by making them accessible to OME managed data, and takes advantage of OME’s features for organized analysis: workflow manager, provenance recording, result reuse, and distributed analysis. The MATLAB Handler is a general method for incorporating legacy MATLAB code with the OME analysis system. We used it to implement a complex image analysis workflow for calculating >800 numerical image descriptors as part of a general image classification technique.
Ilya G. Goldberg