Detection of Auditory Ossicles Using Feature Extraction Method on CT Scan Images

Ear infection, particularly in the middle ear, can lead to precipitation of auditory ossicles─a frequent condition causing marked hearing loss. This study aimed to locate ossicles by extracting features from images obtained by computed tomography (CT) scans. This would help doctors in the rapid and precise diagnosis of patients with infections in the auditory ossicles. The research methodology encompassed extracting features from radiological CT-Scan medical images, for which a pre-processing phase was warranted to enhance image fidelity. The process encompassed contrast stretching, cropping, resizing, filtering, and optimization of contrast to yield images of sharpened clarity specific to the topic of the study. After that, the paper involved object detection through the analysis of the Regions of Interest after marking particular portions in the images. It was followed by shape extraction to find the patterns that define the boundary of the infection. This helps in the differentiation between normal and eroded ossicular conditions. This work attains a rate of 92% in detection accuracy, offers an idea about the extension of auditory ossicles, and much more explicit images are achieved than the original ones. Such findings can help improve the diagnosis skills of the otolaryngology specialist and, therefore, aid better results in patient management. The future work can develop real-time processing capabilities to assist otolaryngologists during surgery or in clinical settings, providing immediate feedback and enhancing decision-making processes.