Research on Object Detection in Resource-Constrained Devices in Edge Computing Scenarios

With the growth of the Internet of Things (IoT) and mobile devices, edge computing has emerged. However, constraints such as memory, storage and power consumption pose challenges for object detection in edge computing scenarios. This paper reviews traditional object detection techniques as well as deep learning models for object detection and introduces two model architectures, ResNet-18 and SqueezeNet, and compares these models through experiments based on the CIFAR-100 dataset. The results show that ResNet-18 has higher accuracy but requires more resources, while SqueezeNet has fewer parameters, faster training time, and smaller memory requirements, making it more suitable for resource-constrained edge computing environments. This study provides a reference for object detection in edge computing scenarios and highlights the need for further research into optimisation methods.