Generalized post-training quantization for medical image segmentation foundation model.
Authors
Affiliations (18)
Affiliations (18)
- Southwest Jiaotong University, School of Computing and Artificial Intelligence, Chengdu, 611756, Sichuan, China; University at Albany, State University of New York, College of Integrated Health Sciences, Albany, 12222, NY, USA. Electronic address: [email protected].
- University at Albany, State University of New York, Department of Mathematics and Statistics, Albany, 12222, NY, USA. Electronic address: [email protected].
- University at Albany, State University of New York, Department of Mathematics and Statistics, Albany, 12222, NY, USA. Electronic address: [email protected].
- University at Albany, State University of New York, College of Integrated Health Sciences, Albany, 12222, NY, USA. Electronic address: [email protected].
- University at Albany, State University of New York, Department of Mathematics and Statistics, Albany, 12222, NY, USA. Electronic address: [email protected].
- University at Albany, State University of New York, Department of Mathematics and Statistics, Albany, 12222, NY, USA. Electronic address: [email protected].
- IBM T. J. Watson Research Center, Yorktown Heights, 10598, NY, USA. Electronic address: [email protected].
- Southwest Jiaotong University, School of Computing and Artificial Intelligence, Chengdu, 611756, Sichuan, China. Electronic address: [email protected].
- Purdue University, School of Applied and Creative Computing, West Lafayette, 47907, IN, USA. Electronic address: [email protected].
- Chengdu University of Information Technology, School of Computer Science, Chengdu, 610225, Sichuan, China. Electronic address: [email protected].
- Chengdu University of Information Technology, School of Computer Science, Chengdu, 610225, Sichuan, China. Electronic address: [email protected].
- Chengdu University of Information Technology, School of Computer Science, Chengdu, 610225, Sichuan, China. Electronic address: [email protected].
- Rensselaer Polytechnic Institute, Troy, 12180, NY, USA. Electronic address: [email protected].
- University at Albany, State University of New York, Department of Computer Sciences, Albany, 12222, NY, USA. Electronic address: [email protected].
- Chengdu University of Information Technology, School of Computer Science, Chengdu, 610225, Sichuan, China. Electronic address: [email protected].
- University at Albany, State University of New York, Department of Computer Sciences, Albany, 12222, NY, USA; University at Albany, State University of New York, AI Plus Institute, Albany, 12222, NY, USA. Electronic address: [email protected].
- University of North Carolina at Chapel Hill, Department of Biostatistics, Chapel Hill, 27599, NC, USA. Electronic address: [email protected].
- University at Albany, State University of New York, College of Integrated Health Sciences, Albany, 12222, NY, USA; University at Albany, State University of New York, AI Plus Institute, Albany, 12222, NY, USA. Electronic address: [email protected].
Abstract
Medical image segmentation foundation models (MedFMs) perform strongly across diverse imaging modalities, but their large size and computational demands hinder deployment in resource-limited clinical settings. Lightweight fine-tuning is impractical given the high training cost of MedFMs, and the efficiency benefits of integer inference remain underused. Post-training quantization (PTQ) is a promising alternative, yet existing PTQ methods fail on MedFMs because their heterogeneous weight distributions lead to severe accuracy degradation at low bit-widths. To address this challenge, we propose GPTQ-MedFM, a generalized post-training quantization framework tailored for medical foundation models. GPTQ-MedFM standardizes complex weight distributions under an ℓ<sub>∞</sub>-constrained normalization to produce quantization-friendly matrices, and then applies an efficient coordinate-descent solver to obtain high-fidelity low-bit representations. The method adds no extra computation or memory overhead at inference, enabling seamless deployment on medical edge devices. By explicitly modeling and mitigating quantization-induced errors, GPTQ-MedFM achieves state-of-the-art low-bit compression across six medical foundation models and nine imaging modalities - spanning nearly the full range of clinical imaging scenarios - and remains robust even with a single calibration sample. Its broad generalization and minimal calibration cost make GPTQ-MedFM a practical pathway for real-time AI-assisted diagnostics in resource-limited healthcare settings.