Towards Efficient Object Re-Identification with A Novel Cloud-Edge Collaborative Framework

TL;DR

The paper tackles the scalability of object Re-Identification in multi-camera networks by introducing a cloud-edge collaborative inference framework powered by Distribution-aware Correlation Modeling (DaCM). DaCM embeds spatial-temporal correlations into a graph to guide both cloud bandwidth allocation and edge image reordering, enabling efficient and accurate inference with existing ReID backbones. Key contributions include the DaCM architecture, its training objective, an inference strategy that couples cloud-level bandwidth allocation with edge-level re-ranking, and a new set of evaluation protocols. The approach reduces data transmission while improving retrieval accuracy in realistic, bandwidth-constrained deployments, offering a practical path toward scalable ReID systems in real-world surveillance networks.

Abstract

Object re-identification (ReID) is committed to searching for objects of the same identity across cameras, and its real-world deployment is gradually increasing. Current ReID methods assume that the deployed system follows the centralized processing paradigm, i.e., all computations are conducted in the cloud server and edge devices are only used to capture images. As the number of videos experiences a rapid escalation, this paradigm has become impractical due to the finite computational resources in the cloud server. Therefore, the ReID system should be converted to fit in the cloud-edge collaborative processing paradigm, which is crucial to boost its scalability and practicality. However, current works lack relevant research on this important specific issue, making it difficult to adapt them into a cloud-edge framework effectively. In this paper, we propose a cloud-edge collaborative inference framework for ReID systems, aiming to expedite the return of the desired image captured by the camera to the cloud server by learning the spatial-temporal correlations among objects. In the system, a Distribution-aware Correlation Modeling network (DaCM) is particularly proposed to embed the spatial-temporal correlations of the camera network implicitly into a graph structure, and it can be applied 1) in the cloud to regulate the size of the upload window and 2) on the edge device to adjust the sequence of images, respectively. Notably, the proposed DaCM can be seamlessly combined with traditional ReID methods, enabling their application within our proposed edge-cloud collaborative framework. Extensive experiments demonstrate that our method obviously reduces transmission overhead and significantly improves performance.

Figures (6)

• Figure 1: Illustration of the difference between centralized and cloud-edge collaborative patterns for ReID systems.
• Figure 2: The overview of our proposed cloud-edge collaborative inference framework. The DaCM is deployed in both the cloud server and edge devices for adjusting uploading batch size $b^i$ and image order in the uploading sequence. The red solid denotes the data flow enabled by the designed DaCM, and the gray dashed line denotes the previous data flow that can be removed by DaCM.
• Figure 3: The architecture of DaCM network.
• Figure 4: The effects of different values of $\alpha$ (upper) and $\beta$ (lower) for mTN and mAP.
• Figure 5: Effects of various $\gamma_0$ (left) and $\gamma_1$ (right) for mTN.