“What we’re developing at the MedUni Vienna is a tool to assist with better diagnosis when, for example, an unknown anomaly is discovered on a CT scan and it would be useful to compare this finding with similar cases in order to arrive at a diagnosis faster,” explains Georg Langs from the University Department of Radiology and Nuclear Medicine who, in collaboration with Franz Kainberger and René Donner in the Computational Imaging Research laboratory (CIR/www.cir.meduniwien.ac.at), is heading up the project at the MedUni Vienna. The project is being unveiled this week under the name of KHRESMOI (www.khresmoi.eu) at the world’s largest trade fair for information technology, CeBIT, in Hanover.
The researchers from Vienna will particularly be focusing on intelligent search algorithms that will be able to spot similarities in large image databases by recognising patterns. The research project is currently being populated with anonymised results from radiological investigations from the Vienna General Hospital. The MedUni Vienna experts are developing methods and algorithms for the computer-based search of radiology image databases, where the image information itself acts as the basis for identifying the search results. Several thousand data records are currently available. The aim of the project is to highlight the technical feasibility – through to the comparison of subtle diagnostic disease characteristics in large medical image databases.
The algorithms developed in the project are intended to help radiologists in their search for relevant cases in a hospital's image database and support them in their reporting routines. The top 100 matching precedence cases are shown – with research from the image database and cross-references to relevant literature.
The project is also aiming to develop a search engine for a wide audience (“KHRESMOI for everyone”), which will serve as a platform for public health with relevant content that has been checked for quality by experts that can also be accessed by private users to provide answers to their enquiries.
The project brings together twelve partners from Europe whose specialist fields range from linguistics therapy and computer-based image processing to medicine.
Five research clusters at the MedUni Vienna
The study falls within the remit of the Medical Imaging research cluster. There are five research clusters in total. The MedUni Vienna is increasingly focusing on fundamental and clinical research in these areas. The other four research clusters are Immunology, Cancer Research / Oncology, Cardiovascular Medicine and Medical Neurosciences.