Experimental Evaluation of Cloud-Based Facial Emotion Recognition Services

keywords: Affective computing, cloud computing, facial emotion recognition, Software as a Service
The main goal of this paper is to perform an extensive analysis of the accuracy of six selected cloud-based facial emotion recognition services on three facial images datasets. The evaluation was performed on more than 33 000 images depicting eight different emotions. Results show that emotion recognition services show a varying level of accuracy over different types of datasets, having a lower accuracy for images of lower quality, but performing considerably better for images taken in ideal conditions. Based on these results we believe that cloud-based facial emotional recognition services do not have the expected accuracy for some use cases and therefore must be selected with care when developing a system that relies on emotion-based interactions.
reference: Vol. 40, 2021, No. 6, pp. 1295–1321