A Multimodal Machine Learning Framework for Diagnosis of Otitis Media with Effusion Using 3D Wideband Acoustic Immittance

Wideband acoustic immittance (WAI) technology has been known for over a decade, delivering an enhanced diagnosis of middle ear (ME) diseases across a wider frequency range than standard tympanometry. Nevertheless, its clinical usage confronts the limitations of restricted interpretation and insufficient explanation of the WAI outcomes. This paper proposes a multimodal machine learning (MML) approach for classifying ME diseases into normal ear and ear with abnormalalty i.e., otitis media with effusion. The proposed MML model is grounded on the integration of a 3 layered convolutional neural network and a multi-layer perception network. The outcomes exhibited that the proposed MML model surpasses the available methods by achieving 98.27% accuracy for classifying ME diseases using the WAI measurements.