Measurement of Hearing Loss due to Perforated Tympanic Membrane using Image Processing Techniques

摘要

The tympanic membrane (ear drum) is a thin tissue film that is stretched between the outer and middle ear. Sound waves travel from outside the ear, and strike the tympanic membrane resulting in its vibration. These vibrations amplify the sound waves and transmit them to the ossicles (auditory bones). The magnitude of amplification is directly proportional to vibrating area of tympanic membrane. Hence a perforation in this membrane would result in hearing loss. Pure-tone audiometry is the traditional procedure used to detect the amount of hearing loss in a patient. However, it is lengthy and less efficient, as it largely depends on the response of the patient to sound intensity and frequency of pure-tones. We present a relatively more efficient approach to determine hearing loss due to perforated tympanic membrane using image processing techniques. We describe an algorithm that uses unsharp masking to sharpen images of the perforations as well as the tympanic membrane. Then, it converts the image into a binary image using thresholding. A median filter is applied to get rid of the noise component in the image. The ratio of the area of perforation and total area of tympanic membrane will define the percentage of hearing loss. Our approach will eliminate the error introduced due to patient dependency as in the traditional method.