Analysis of Deep Learning Models for Disease Diagnosis in Laryngoscopic NBI Images

The rising trends about disorders related to larynx diseases are alarming. These are mainly caused by issues such as environmental pollution, aging population and more inappropriate use of harmful substances. Such disorders have a very dire effect on the health aspects of people and may lead to graver symptoms. Thus, proper urgent treatment is required for producing timely results in patients to control the impacts of such diseases. Originally developed as an instrument to examine cervical and uterine and epithelial cells for disease diagnosis, contact endoscopy has become a very valuable diagnostic technique for several conditions of the larynx. The terms Contact Micro-Laryngoscopy is used when this technique is employed to examine the larynx. This is a minimal invasive technique, which allows direct real-time visualization of the cellular and vascular anatomy of the superficial layer of the laryngeal mucosa. The idea in this research if the harnessing power of Contact Endoscopy-Narrow Band Imaging together. Using CE-NBI, improves the capability of providing diagnosis, highlights anomalies occurring in the larynx.