ESOD:Edge-based Task Scheduling for Object Detection
This addresses the problem of efficient object detection for mobile users by reducing latency and energy use, but it is incremental as it builds on existing edge computing and scheduling methods.
The paper tackles the challenge of high computation burden for object detection on mobile systems by proposing ESOD, an edge-based task scheduling framework that uses a pre-model to predict and offload tasks to edge servers based on image characteristics, resulting in average reductions of 22.13% in latency and 29.60% in energy consumption, and improving mAP to 45.8.
Object Detection on the mobile system is a challenge in terms of everything. Nowadays, many object detection models have been designed, and most of them concentrate on precision. However, the computation burden of those models on mobile systems is unacceptable. Researchers have designed some lightweight networks for mobiles by sacrificing precision. We present a novel edge-based task scheduling framework for object detection (termed as ESOD). In detail, we train a DNN model (termed as pre-model) to predict which object detection model to use for the coming task and offloads to which edge servers by physical characteristics of the image task (e.g., brightness, saturation). The results show that ESOD can reduce latency and energy consumption by an average of 22.13% and 29.60% and improve the mAP to 45.8(with 0.9 mAP better), respectively, compared with the SOTA DETR model.