Efficient Feature Compression for Machines with Global Statistics Preservation

TL;DR

This paper introduces an efficient feature compression method using Z-score normalization for split-inference AI models, reducing bitrate and improving accuracy, and is integrated into the MPEG FCM standard.

Contribution

The paper presents a novel Z-score normalization-based compression technique that outperforms existing methods in the MPEG FCM standard for split-inference models.

Findings

17.09% average bitrate reduction across tasks

Up to 65.69% bitrate reduction for object tracking

Improved end-task accuracy with less data transfer

Abstract

The split-inference paradigm divides an artificial intelligence (AI) model into two parts. This necessitates the transfer of intermediate feature data between the two halves. Here, effective compression of the feature data becomes vital. In this paper, we employ Z-score normalization to efficiently recover the compressed feature data at the decoder side. To examine the efficacy of our method, the proposed method is integrated into the latest Feature Coding for Machines (FCM) codec standard under development by the Moving Picture Experts Group (MPEG). Our method supersedes the existing scaling method used by the current standard under development. It both reduces the overhead bits and improves the end-task accuracy. To further reduce the overhead in certain circumstances, we also propose a simplified method. Experiments show that using our proposed method shows 17.09% reduction in bitrate on average across different tasks and up to 65.69% for object tracking without sacrificing the task accuracy.