Weighted Fuzzy ARTMAP for Osteoporosis Detection

Osteoporotic is a health burden worldwide, resulting in reduction of physical activity, increased risk of mortality, and incremental medical cost. Mandibular trabecular patterns analyzed on dental panoramic radiographs have been widely studied for identifying postmenopausal women with low skeletal bone mineral density (BMD). In this paper we proposed a new method for detecting osteoporosis using Weighted Fuzzy ARTMAP from the features measured in dental panoramic radiographs. The method developed an activation match function by integrating Simplified fuzzy ARTMAP and symmetric Fuzzy ART. Fourier method and segmentation processing were applied for obtaining features of a radiograph in frequency and spatial domain. We also introduced an additional weighted parameter based on pheromone to distinguish clusters based on the amount of its member. The experimental results for osteoporosis detection show that the new method achieved accuracy of 87.88%, sensitivity of 93.33%, and specificity of 83.33%.