Adaptive Hybrid Clustering for Automated Teeth Segmentation Framework with Improved Eurasian Oystercatcher-Based Weighted Fused Features

Selecting a proper region of interest with a high success rate and more accuracy is a highly challenging task in dental diagnosis systems. Dental professionals are struggling to analyze the usage of X-ray images. Therefore, this research concentrates on developing an efficient teeth segmentation model to address the challenges in dental caries diagnosis. The dental images are collected from the standard datasets. Further, pre-processing is performed utilizing Contrast Limited Adaptive Histogram Equalization (CLAHE), bilateral filtering and histogram equalization to get the pre-processed images. Subsequently, the weighted fused features are carried out by fusing the features from Red Green Blue (RGB), Pixel Intensity Edge Entropy (PEE), Local Binary Pattern (LBP) and gradient feature extraction techniques. Here, weight optimization takes place by Modified Mussel Length-based Eurasian Oystercatcher Optimizer (MML-EOO) for enhancing the segmentation with significant features. Finally, segmentation is performed by an Adaptive Hybrid Clustering (AHC) with the combination of Fuzzy C-Means (FCM) and K-Means Clustering (KMC) to make accurate segmentation with weighted fused features. The parameters in FCM and KMC are tuned with the same MML-EOO algorithm to enhance the overall segmentation efficiency. The experimental results have shown that the developed teeth segmentation model obtains higher efficiency than the existing techniques.