Modular Quantization-Aware Training for 6D Object Pose Estimation

TL;DR

This paper introduces Modular Quantization-Aware Training (MQAT), a novel adaptive mixed-precision quantization method that significantly compresses 6D pose estimation models while improving accuracy, suitable for resource-limited edge applications.

Contribution

MQAT is a new modular, mixed-precision quantization-aware training strategy that outperforms existing methods in compressing 6D pose estimation networks with accuracy gains.

Findings

Achieves over 7% accuracy improvement over full-precision models.

Reduces model size by at least 4 times.

Demonstrates robustness across datasets, architectures, and quantization algorithms.

Abstract

Edge applications, such as collaborative robotics and spacecraft rendezvous, demand efficient 6D object pose estimation on resource-constrained embedded platforms. Existing 6D pose estimation networks are often too large for such deployments, necessitating compression while maintaining reliable performance. To address this challenge, we introduce Modular Quantization-Aware Training (MQAT), an adaptive and mixed-precision quantization-aware training strategy that exploits the modular structure of modern 6D pose estimation architectures. MQAT guides a systematic gradated modular quantization sequence and determines module-specific bit precisions, leading to quantized models that outperform those produced by state-of-the-art uniform and mixed-precision quantization techniques. Our experiments showcase the generality of MQAT across datasets, architectures, and quantization algorithms. Remarkably, MQAT-trained quantized models achieve a significant accuracy boost (>7%) over the baseline full-precision network while reducing model size by a factor of 4x or more. Our project website is at: https://saqibjaved1.github.io/MQAT_/