Biodata Mining Group


The automated analysis of digital underwater images requires complex approaches of machine learning. Central problems are the detection of objects as well as the following semantic classification of plants, animals or mineral ressources.

An interest in data mining and/or image analysis is required and skills in efficient coding are beneficial but not mandatory. We target high-throughput solutions and the hundreds of thousands of images to be analysed require that systems are developed in C++ with the help of our own machine learning library and/or OpenCV. Computational speedup can be achieved through GPU enabled algorithms or parallelization on the CeBiTec compute cluster.

Big data challenges the ways we think about data, analyze data and handle data. We in the Biodata Mining Group follow a collaborative strategy and develop web-based software tools and libraries for cooperative data analysis. Tools are designed, developed and operated to handle multi-dimensional data from various fields like marine environments or tissue samples.

The tool development includes novel visualization strategies, efficient methods of data retrieval, fusion of different data types and much else.

We rely on current Internet technologies (PHP/Python, MySQL, HTML5, CSS3, JavaScript, JSON and JSON-RPC) and include libraries as Bootstrap and jQuery.