Automatic diagnosis of Hypoacusia with Associative Memories

Abstract

Hypoacusia 1 is the reduction in hearing ability. An early diagnose could avoid the complete loss of the sense of hearing. We propose a modification of modified Johnson-Möbius together with a tool of Artificial Intelligence to diagnose hearing losing. The modified Johnson-Möbius has been showed suitable results when it was used with Alpha-Beta associative memories that deal with binary numbers. Now, we modified this code to apply it to Morphological associative memories whose set of numbers is real-type. Based on the improved results of Alpha-Beta memories with this codification, we applied the modification of the code to improve the performance of Morphological memories with feature selection. The results are suitable to implement an automatic system for hypoacusia diagnosis.