Computer Vision on Edge: Using Edge Impulse and Deep Learning Models on Raspberry Pi

Some of the applications in the field of machine learning and computer vision may require a higher level of privacy and security, better cost efficiency, offline operation, and greater scalability and efficiency. This can be ensured by building machine learning and computer vision on edge devices. This paper presents solutions in the field of computer vision using deep learning models implemented in Edge Impulse and then deployed for use on the Raspberry Pi. Deep neural network models were used, YOLO (You Only Look Once), SSD (Single Shot MultiBox Detector), MobileNet, and EfficientDet.