The FER-Based Module embedded in SenGAware Tool: A Group Emotion Awareness Perspective in CSCL

Emotions play a crucial role in collaboration, and timely and effective emotion awareness is particularly important in the collaborative process. In recent years, emotional awareness has gradually evolved from subjective measurements of explicit emotional information to the recognition and analysis of implicit emotional information. This paper focuses on the implementation and application of a facial expression recognition module embedded in the group emotion awareness tool, SenGAware. It technically leverages the Face.evoLVe facial recognition library and the RESNET50 network to develop a facial recognition model, while utilizing the PaddleX library and the ResNet50_vd_ssld network to establish a facial expression recognition model. Ultimately, it primarily achieves the awareness of emotional states of team members during collaboration, and through the emotional data generated, subsequently produces emotional reports through comprehensive LLM analysis. It not only enhances the precision of emotion awareness in the facial expression recognition component of the SenGAware tool, but also facilitates improved emotional interaction and collaborative capabilities among team members, presenting a novel research angle for the development of implicit emotion awareness technology in CSCL environments.