Abstract—Infections of the ear, especially of the middle ear, which may result in the loss of the auditory bones, most commonly affect the hearing ability of man. This is because, in the middle part of a human’s ear, these infections mainly result in the loss of three major auditory ossicles: malleus, incus, and stapes, which transmit sound vibrations from the tympanic membrane to the cochlea. Up to 60% of the hearing can be lost, with a harmful impact on human communication and interaction. Even though the examination is done through Computed Tomography (CT) scans, proper image interpretation is still essential for diagnoses. This project will research an advanced feature extraction method from the CT scan image to diagnose auditory bone erosion quickly. The research methodology of this project will be performed in several steps, which start from the input images and later include the pre-processing steps such as contrast stretching, cropping, re-sizing, filtering, contrast adjustment, and histogram equalization. The processing stage included object detection using the Region of Interest (ROI) method and shape extraction based on four parameters: area, perimeter, metric, and eccentricity. A novel algorithm had been designed purposely for shape extraction based on area parameters and was named the Chain Code algorithm; the method is named the MCACC method. Then, Experimental results demonstrate that from epoch 1, iteration 1, to epoch 100, iteration 100, the model achieves an accuracy rate of 97.37%. These findings highlight the efficacy of the MCACC algorithm in detecting erosion areas in auditory ossicles, providing a robust tool for medical imaging analysis. The detailed information regarding bone erosion with more explicit images obtained in this work significantly supports more advanced diagnosis and treatment in otolaryngology, thus enhances both patient care and communication capability.
 
Keywords—feature extraction, extraction method, CT-scan image, ossicles erosion, auditory, Convolutional Neural Network (CNN), Chain Code (CC)


Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.