Published Paper Details:

The integration of artificial intelligence with Internet of Things (IoT) devices has opened new frontiers in real-time human-computer interaction, especially in emotion-aware systems. However, traditional facial emotion recognition (FER) models demand high computational power, making them unsuitable for resource-constrained edge devices. This paper presents a lightweight facial emotion recognition model optimized for deployment on a Raspberry Pi, enabling real-time emotion detection at the edge without relying on cloud servers. The proposed system utilizes a compact convolutional neural network (CNN) trained on the FER-2013 dataset, which is converted to TensorFlow Lite format for efficient inference. Facial regions are detected using OpenCV's Haar Cascade classifier, followed by on-device emotion classification into key categories such as happy, sad, angry, surprised, and neutral. The system also integrates with Adafruit IO to transmit emotion data over Wi-Fi for remote monitoring and IoT-based applications. Performance evaluation demonstrates low latency, reduced power consumption, and high accuracy in real-time conditions, confirming the suitability of Raspberry Pi for edge-deployed AI tasks. The proposed model has potential applications in smart homes, emotion-aware healthcare monitoring, and intelligent human-machine interfaces.