Feature Coding for Scalable Machine Vision

TL;DR

This paper introduces a feature coding standard and test model that significantly reduces bandwidth for edge-cloud machine vision inference, enabling efficient, privacy-preserving deployment.

Contribution

It presents the design and performance evaluation of the FCTM, demonstrating substantial bitrate reduction while maintaining accuracy in scalable machine vision.

Findings

Average bitrate reduction of 85.14% across tasks

Preserves accuracy despite compression

Enables bandwidth-efficient edge-cloud inference

Abstract

Deep neural networks (DNNs) drive modern machine vision but are challenging to deploy on edge devices due to high compute demands. Traditional approaches-running the full model on-device or offloading to the cloud face trade-offs in latency, bandwidth, and privacy. Splitting the inference workload between the edge and the cloud offers a balanced solution, but transmitting intermediate features to enable such splitting introduces new bandwidth challenges. To address this, the Moving Picture Experts Group (MPEG) initiated the Feature Coding for Machines (FCM) standard, establishing a bitstream syntax and codec pipeline tailored for compressing intermediate features. This paper presents the design and performance of the Feature Coding Test Model (FCTM), showing significant bitrate reductions-averaging 85.14%-across multiple vision tasks while preserving accuracy. FCM offers a scalable path for efficient and interoperable deployment of intelligent features in bandwidth-limited and privacy-sensitive consumer applications.