Analyzing Face and Speech Recognition to Create Automatic Information for Usability Evaluation

摘要

Observe users perform their tasks in a software is an important way to performing usability evaluation due to the reason that provides real data about the interaction between user and system. Filming and verbalization are very used techniques and they must be a concern for all designers. However, the needs of reviewing all registered data manually became these techniques slow and difficult. This paper presents an approach that uses face recognition and speech processing to generate relevant information about a system under test such as what moments the user had specific reaction and which ones. The ErgoSV software supported the experiments that were performed using an e-commerce website. The results showed that this approach allows the evaluator identify interfaces with usability problems easily and quickly as well as present information using percentages that supported the evaluator making decision.