Implementing UX/UI in a real-time speech recognition and translation application in Android and a clinical study

Abstract: According to WHO(World Health Organization) there are over 1,5 billion people living with hearing loss. Furthermore, HRF(Hörselskadades riksförbund) believe that ‘hidden hearing loss’ can be a much larger scale of the population than what was believed. There are two parts in this project taking place parallel. The first part is a clinical study which purpose is to compute more samples from the hearing test application made by Robin Söderberg and Amin Saremi which tests hearing loss and hidden hearing loss with a puretone audiometry test and a temporal masking test. Furthermore, the purpose is to detect if there are people with hidden hearing loss, check if the application gives correct results and compare the results to similar clinical studies. The second part is implementing a UX/UI design for a real-time speech recognition and translation application in Android using Javaand XML. The purpose is for the design to be user friendly for the end user group which are people with hearing impairment and second language users. The first part correlates to the second part through the applications end user group which are studied in the clinical study.The clinical study is estimated to be performed on 10 participants with a mixture of ages and an approximately 50/50 percentage of both female and male participants. A script was made to inform all the participants the same information and a question template was made to riddle out any factors which could affect the test results such as fatigue, temporary hearing loss and hearing loss. The UX/UI design uses View and a part of Controller from the MVC(Model-View-Controller) method when implementing the design. The Design Thinking method is used to construct a user-friendly design for the end user group. Results from the clinical study revealed expected results from the 10 normal hearing participants on the audiometry test and temporal masking test. The 10 normal hearing participants were 5 males and 5 females with an average value of 30,70 years old and a standard deviation of 6,02 years. Findings showed a distinct result could indicate hidden hearing loss. The results are similar to an extent but are more accurate in comparison to similar clinical. A possibility for this difference could be due to the bigger sample size, the presence of the question template or some other factors. When examining the similarities to similar clinical studies the application gives correct results for normal hearing participants. Results from the UX/UI design revealed the design for the real-time speech recognition and translation application using the Design Thinking method. All the functions have the end user group in mind in theory based on the methods used. A test with real end users to test the application was not able to be executed due to time limitations and therefore it is difficult to conclude if the end user group are satisfied with the application and find it helpful. Conclusions from the clinical study are that the hearing test application could possibly detect people with hidden hearing loss. The results from the application can give more accurate results for normal hearing participants depending on the execution or factors during testing. Furthermore, the application gives correct results for normal hearing participants. More testing needs to be done on a bigger sample size to give more accurate results and possibly find more people with hidden hearing loss. Conclusions from the UX/UI design revealed the design on the real time speech recognition and translation application could be user friendly in theory but needs to be tested on the end user group to conclude if it is user friendly in praxis and if the end user group find the application helpful and satisfying.