• September 2024: Paper accepted in ACM/IEEE Symposium on Edge Computing (SEC 2024)

The paper titled “Optimizing Edge Offloading Decisions for Object Detection” co-authored with my student Jiaming Qiu, as well as Ruiqi Wang, Brooks Hu, and Chenyang Lu was just accepted for presentation at the ACM/IEEE Symposium on Edge Computing (SEC 2024).

The paper addresses the problem of object detection in an edge computing setting, where a (weak) detector is available on-board cameras capturing images of objects to facilitate distributed decisions, but is supplemented by a shared (stronger) edge detector accessible across a network and available when local decisions are uncertain. Resource constraints, e.g., network or edge detector capacity, however, limit the number of images that can be offloaded to the edge. The paper’s goal is to identify which images to offload to maximize overall detection accuracy.
Its main contributions are a metric and a method for estimating it, which enable efficient online offloading decisions. The approach is computationally frugal enough to run on embedded devices, and empirical findings indicate that it outperforms existing alternatives in improving detection accuracy even when the fraction of offloaded images is small.