Computer Aided Diagnosis Based Hand Thermal Image Analysis: A Potential Tool for the Evaluation of Rheumatoid Arthritis
The aim of the study was to evaluate the potential of the thermogram in diagnosing rheumatoid arthritis and to compare the implementation of k-means algorithm and fuzzy c means algorithm using a computer aided diagnostic tool for classification of rheumatoid arthritis (RA) and normal based on the feature extracted from the segmented thermal image. The skin surface temperature measurement, thermal image segmentation based on k-means and fuzzy c-means algorithm and the feature extraction were used in this study. The average skin surface temperature measured at the second meta carpo-phalangeal (MCP) and MCP3 in RA patients (35.40 ± 0.6 °C and 35.52 ± 0.7 °C, respectively) were significantly higher (p < 0.01) than those measured in normal subjects (33.66 ± 0.2 °C and 33.74 ± 0.2 °C, respectively). The mean difference in temperature between RA patients and healthy controls in the MCP2 and MCP3 region was found to be 1.74 and 1.78 °C respectively. The receiver operating characteristics (ROC) curve depicted a sensitivity of 86.6% and specificity of 79% achieved in the MCP region of the hand thermal image. Thermal image segmentation using the k-means algorithm provided better segmentation results compared to the fuzzy c-means algorithm in diagnosing the disease. Therefore, the computer aided diagnostic based thermography method could be used as a validated quantification method for interpreting and evaluating arthritis.