Feature Compression for Machines with Range-Based Channel Truncation and Frame Packing

TL;DR

This paper introduces a novel feature compression method using range-based channel truncation and frame packing to improve bandwidth efficiency in split computing for neural network inference, achieving over 10% rate reduction.

Contribution

It proposes a new channel truncation and packing technique that preserves relevant features and maintains task performance in feature coding for split neural network inference.

Findings

Achieved an average of 10.59% rate reduction at constant accuracy.

Effective preservation of relevant feature channels during compression.

Improved bandwidth efficiency for feature transmission in split computing.

Abstract

This paper proposes a method that enhances the compression performance of the current model under development for the upcoming MPEG standard on Feature Coding for Machines (FCM). This standard aims at providing inter-operable compressed bitstreams of features in the context of split computing, i.e., when the inference of a large computer vision neural-network (NN)-based model is split between two devices. Intermediate features can consist of multiple 3D tensors that can be reduced and entropy coded to limit the required bandwidth of such transmission. In the envisioned design for the MPEG-FCM standard, intermediate feature tensors may be reduced using Neural layers before being converted into 2D video frames that can be coded using existing video compression standards. This paper introduces an additional channel truncation and packing method which enables the system to preserve the relevant channels, depending on the statistics of the features at inference time, while preserving the computer vision task performance at the receiver. Implemented within the MPEG-FCM test model, the proposed method yields an average reduction in rate by 10.59% for a given accuracy on multiple computer vision tasks and datasets.