What question did this study set out to answer?

The aim is to introduce MedPTQ, a pipeline that improves inference efficiency in 3D medical image segmentation without losing accuracy.

February 21, 2026Open Access

MedPTQ: a practical pipeline for real post-training quantization in 3D medical image segmentation

Key Points

The aim is to introduce MedPTQ, a pipeline that improves inference efficiency in 3D medical image segmentation without losing accuracy.
Developed a post-training quantization pipeline called MedPTQ
Validated across various AI architectures, including CNNs and transformers
Tested on diverse medical imaging datasets from multiple hospitals
MedPTQ achieves INT8 inference, reducing model size and computational demands
Segmentation accuracy remains comparable to full-precision baselines
Demonstrates generalizability across different imaging modalities and body regions

Abstract

We have introduced MedPTQ, a real post-training quantization pipeline that delivers INT8 inference for SOTA 3D artificial intelligence (AI) models in medical imaging segmentation. MedPTQ effectively reduces real-world model size, computational requirements, and inference latency without compromising segmentation accuracy on modern GPUs, as evidenced by mDSC comparable to full-precision baselines. We validate MedPTQ across a diverse set of AI architectures, ranging from convolutional-neural-network-based to transformer-based models, and a wide variety of medical imaging datasets. These datasets are collected from multiple hospitals with distinct imaging protocols, cover different body regions (such as the brain, abdomen, or full body), and include multiple imaging modalities computed tomography (CT) and magnetic resonance imaging (MRI). Collectively, these results highlight our MedPTQ's strong generalizability and adaptability for a broad spectrum of medical imaging tasks.

KI fragen

Bookmark

View Full Paper