Information Fusion of Image Analysis, Video Object Tracking, and Data Mining of Biological Images using the Open Source MATLAB Toolbox Gait-CAD

Abstract

Automated imaging has become a commonplace and widespread technique for researchers aiming to increase both biological and medical knowledge. Systematic high-throughput screening approaches produce a vast amount of data that needs to be quantified automatically. To address this problem, we present an extended version of the open-source MATLAB toolbox Gait-CAD providing integrated tools for automated image analysis, video object tracking and data mining. Gait-CAD offers a convenient graphical user interface (GUI) and is shipped with a great selection of predefined, customizable plugins for both image analysis and data mining. The plugin-based architecture and templates for customized tools provide easy expandability in order to develop comprehensive data-analysis pipelines. Process automation via batch-files and macro recording functionality enables the handling of large datasets like multi-dimensional 2D or 3D images and videos. The scope of the presented tools ranges from automated high-throughput toxicity testing in zebrafish embryos to cellular analysis tasks in developmental biology. In both examples, the toolbox is successfully applied for pre-processing, normalization, segmentation and tracking of spatio-temporal microscopy images, as well as for subsequent data mining and report generation. As automatically acquired images tend to differ in each recording, researchers can significantly accelerate parameter adjustments, process automation and result visualization by using the presented software. The toolbox is not limited to these applications, but they already reveal the great potential of the extended Gait-CAD release. The presented toolbox is a powerful instrument for data analysis in life sciences. A user-friendly GUI provides functionality to create sophisticated approaches even for users with limited programming knowledge.