A computer-aided decision support system for diagnostics of laryngeal diseases

To develop methods and a computer-aided system for automated laryngeal disease diagnostics based on information obtained from multiple sources.

In addition to the data obtained from one particular patient, previous experience gained from many previous patients plays a very important role in the decision making process of a physician. However, the physician interpreting the available data from a particular patient may have limited knowledge and experience in analysis of the data. In such a situation, a decision support system for automated analysis and interpretation of medical data is of great value. Recent developments in this area have shown that physicians benefit from the advice of decision support systems in terms of increased reliability of the analysis, and decreased intra- and inter-observer variability. The visual appearance of vocal cords, a patient's voice and inquiry data are the main information sources usually used in the diagnostics of laryngeal diseases of humans. Colour, geometrical shape, roughness and irregularity are the main features used by a physician to characterise the visual appearance of vocal cords. The result of such an analysis is a verbal characterisation, which is rather subjective and to a great extent depends on the physician's experience. The subjectivity degree increases even more when information available from all the aforementioned sources needs to be integrated to make a conclusion. This project aims to create a computer-aided decision support system for automated diagnostics of laryngeal diseases of humans. The system will provide a method for obtaining objective measurements of features characterising the visual appearance, the patient's voice, the verbal information obtained from experts and other information. A systematic database - a base of images, acoustic signals, verbal and other information which can prove useful for the analysis will also be created. It is expected that the system will be a very valuable tool in everyday health care especially when lacking knowledge or experience, in education processes, and in scientific research when establishing different tendencies and dependencies. To achieve the goals, there will be development of methods and tools for colour image analysis and classification based on the integrated use of colour, texture and geometrical information obtained from the global as well as local analysis. Advanced methods and tools for extracting informative features from all the available information characterising a patient's voice as well as methods for formalising verbal information obtained from experts will be devised. Advanced algorithms for feature selection, information fusion and interpretation of the decisions taken by the system will also be developed. Keywords: laryngeal diseases, image analysis,pattern recognition.
Project ID: 
3 681
Start date: 
Project Duration: 
Project costs: 
320 000.00€
Technological Area: 
Imaging, Image Processing, Pattern Recognition
Market Area: 
Hospital and other institutional management (including management services and leasing)

Raising the productivity and competitiveness of European businesses through technology. Boosting national economies on the international market, and strengthening the basis for sustainable prosperity and employment.