Vasculature Detection in Tympanic Membrane

Tympanic membrane, an essential part of the human ear whose function is to help in hearing. In this paper, novelty-based characterization of the tympanic membrane using its vasculature is proposed. In part (1) digital image processing techniques were used for the detection of the blood vessels. In this green channel extraction followed by pre-processing techniques like CLAHE and OTSU thresholding were performed. Vessel detection was done by using Morphological Operation combined with Canny Edge Detection. In part (2) Histogram and pixel-based techniques were used for quantization of blood vessels. Result analysis was done using percentage of vasculature and speed of the algorithm used in the finding the percentage of vasculature. Comparison between Sobel edge detection and Canny edge detection was performed to choose the best method.

I. Introduction

Tympanic membrane is a thin, oval, and translucent structure that separates the external ear from the middle ear. It is approximately O.lmm thick, 8-10mm in diameter and lies at an angle of 45 degree to the horizontal axis. The sound waves traveling from the outer ear strike the tympanic membrane causing it to vibrate. This vibration is then transmitted by the tympanic membrane into the inner ear. The tympanic membrane has three layers, outer, inner, and middle layer. The outermost layer of the tympanic membrane is a continuation of the ear canal skin. The innermost layer is in continuation of the mucous membrane lining the middle ear and the middle layer consists of radial and circular fibers giving the tympanic membrane its stiffness and tension. The internal surface and the external surface of the tympanic membrane are convex and concave respectively.