Detection of Different Types of Ear Diseases in Infants Using Deep Learning for Early Treatment

The loss or inability of auditory sense is the most common sensory organ deficiency. Out of every 1000 new-borns, 5–6 are deaf or hard of hearing. They will not be detected until they are two or three years old, after which permanent trauma would have happened. One of the most prominent reasons for deafness in infants is due to ear diseases. Early detection of ear diseases is the top priority to prevent deafness in infants. This research work has developed and applied modified deep learning architecture such as AlexNet, mini GoogLeNet, and LeNet and compared the training by the three architectures. Of the three, AlexNet architecture detected and identified the presence of different ear ailments such as acute otitis media, glue ear, safe CSOM, otomycosis, other infections and also checked whether the ear is normal, with high accuracy.