Optimized-CNN enabled Facial Emotion Recognition within Collaborative Edge Computing

Emotion recognition as a technology to correctly understand individual emotions could provide guidance for psychological adjustment. Therefore, it is necessary to develop accurate and effective emotion recognition models and algorithms. With the emerging technologies of artificial intelligence and the popularity of video surveillance, the use of neural networks for facial emotion recognition has been proven to be more efficient than traditional methods. However, it still has room for improvement in the delay and robustness of simple neural networks. To address this issue, an optimized convolutional neural network (CNN) enabled facial emotion recognition within collaborative edge computing is investigated in this paper. Within the paradigm of collaborative edge computing, the Raspberry Pi 3B+ acting as the edge computing node is employed to deal with the related issues of multi-view facial emotion recognition. The optimized CNN models deployed on the edge nodes are adopted to collaboratively extract facial features from images and conduct the classification to achieve facial emotion recognition. In terms of the recognition results, the corresponding appropriate psychological adjustment strategies would be pushed to the target. Through using the OpenCV library to implement a prototype on Raspberry Pi 3B+, the experimental results ultimately have been shown to demonstrate the efficiency of the investigations.