4.1. Model Performance Analysis
The experimental results show that the performance of the model for detection-based perception tasks at each layer either with GPU or non-GPU-based devices resulted in the same performance. It is evident from the results that the performance values of metrics such as precision, recall, [email protected], and [email protected]:0.95 are the same for all object classes. For instance, the performance metric values, i.e., precision (0.84), recall (0.90), [email protected] (0.93), and [email protected]:0.95 (0.72) for detecting a car object will remain the same regardless of deploying the same model over any type of device and across any layer of the framework. The main reason behind such behavior of the deployed model is its performance in object detection over the same set of data, using the same weights of the fine-tuned YOLOv8s model. Therefore, we…
