Pattern recognition of laser transillumination images for diagnosis of finger joints diseases

Inflammatory processes of interphalangeal joints very often lead to disability and decreases the quality of life. The early diagnosis, especially in case of rheumatoid arthritis, is essential for proper treatment and monitoring the progress of illness. Transillumination is one of the oldest modalities once used for medical diagnosis. We have designed and constructed the experimental diagnosis set-up consisted of He-Ne laser as a light source, special holder for patient finger, CCD camera and image processing system for classification of images. The system was tested on ill, as well as on healthy persons. For classification the optoelectronic pattern recognition methods were applied. We have proved that transillumination is suitable for diagnosis, providing that a proper recognition procedure will be chosen. It was demonstrated the image processing and classification enables differentiation of healthy and ill individuals with 95% accuracy and monitoring the therapy progress, as well.