ErgoSV: An environment to support usability evaluation using face and speech recognition

Abstract

Usability test is a group of activities that should be performed by all designers in order to identify interaction problems. Filming and Verbalization are two techniques widely used due to the reason that they provide real information about the software interaction capacity. Filming is performed using one or several cameras and the verbalization is done encouraging the participant to verbalize what he/she is thinking about the software. Both techniques register the data in video and audio files to be analyzed forward. Although these techniques has been widely used, the analysis process is considered slow, difficult and expensive because the evaluator may need to review all the data registered from the first second until the end of the test to identify possible usability problems and this task could take from 2x to 10x the test time. This paper presents the ErgoSV Software, a tool to support usability evaluation test using speech processing that recognize specific keywords pronounced by the participants and face images processed during the test. These data are used to provide organized and relevant information to support the data analysis and the identification of interfaces with possible usability problems. Experiments performed in three different softwares presented that this tool reduced the time of analysis to 1,5 times the test time considering the keywords as the main data. This research is supported by FAPESP. © 2014 Springer International Publishing.