Efficient Feature Compression for Edge-Cloud Systems

TL;DR

This paper introduces a new training strategy and encoder architecture for edge-cloud image classification that optimizes the trade-off between bit rate, accuracy, and encoding complexity, improving performance across various settings.

Contribution

It presents a scalable feature coding system that enhances the rate-accuracy-complexity trade-off in edge-cloud systems, outperforming previous methods.

Findings

Outperforms previous methods in RAC performance

Scalable design for different edge device resources

Consistent improvements across various settings

Abstract

Optimizing computation in an edge-cloud system is an important yet challenging problem. In this paper, we consider a three-way trade-off between bit rate, classification accuracy, and encoding complexity in an edge-cloud image classification system. Our method includes a new training strategy and an efficient encoder architecture to improve the rate-accuracy performance. Our design can also be easily scaled according to different computation resources on the edge device, taking a step towards achieving a rate-accuracy-complexity (RAC) trade-off. Under various settings, our feature coding system consistently outperforms previous methods in terms of the RAC performance.