Automated thermal image segmentation of knee rheumatoid arthritis

• Published in: 2016 International Conference on Communication and Signal Processing (ICCSP) pp. 0535 - 0539
• Main Authors: Suma, A. B., Snekhalatha, U., Rajalakshmi, T.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.04.2016
• Subjects: Algorithm design and analysis; Arthritis; Computer-aided diagnosis (CAD); Disease activity score 28 (DAS28); Erythrocyte sedimentation rate (ESR) rate; Health Assessment Questionnaire (HAQ); Image color analysis; Image segmentation; Imaging; Magnetic resonance imaging (MRI); Manuals; Rheumatoid arthritis (RA); Rheumatoid factor (RF factor)
• DOI: 10.1109/ICCSP.2016.7754195

Rheumatoid arthritis (RA) is one of the systemic autoimmune disorders characterized by chronic inflammation in multiple joints eventually leading to bone erosion and cartilage damage. The main symptoms are (i) morning stiffness (ii) joint swelling (iii) inflammation (iv) neovascularisation (v) cell infiltration and (vi) synovial hyperplasia. Early diagnosis helps in the effective treatment of the disease, however, in the early stages there may not be apparent joint swelling and often present negative radiographs. The common diagnostic methods include Magnetic Resonance (MR) imaging and clinical and health assessment questionnaire. However, MRI is quite expensive; hence, there is a need for cost-effective and safer technique for the diagnosis of RA. Therefore, this study focuses on the application of non-invasive, radiation-free, cost economic technique namely thermography in the diagnosis of early stage RA. The main aim of this paper is to compare various segmentation algorithms, namely manual, colour and k-means image segmentation and to find out the best suitable segmentation algorithm for thermal images. The hot spot region is extracted using three different image segmentation algorithms and compared with the normal thermograph to determine the effective segmentation algorithm in the detection of early stage RA.