cs.CVNov 4, 2026

PTQ4SAM3: Semantics-Preserving Post-Training Quantization for Segment Anything Model 3

Authors: Xi Xiao, Yunbei Zhang, Lin Zhao, Janet Wang, Xinru Jiang, Yanshu Li, Chenrui Ma, Pan Wang, 

Organizations: University of Alabama at Birmingham, USA · Tulane University, USA · Northeastern University, USA · University of Texas at Austin, USA · 

Abstract

Segment Anything Model 3 (SAM 3) can segment objects from text prompts and track them across video, but its memory and computation costs make deployment difficult. Post-training quantization (PTQ) reduces these costs by using fewer bits. However, a quantized model can select the wrong object or lose track of its target across frames. We call this prompt-alignment drift. Controlled tests show that perturbing low-variance feature directions strongly affects target selection. We propose Semantics-Preserving Quantization (SPQ), which gives more calibration weight to concept-sensitive parameters, preserves residual features, and corrects the mean error in video memory. Fixed text concepts guide calibration without task annotations. All three components operate offline and add no inference operators. With a W4A8 image encoder, SPQ improves RefCOCO [email protected] from 42.1 to 72.5 over PTQ4SAM while retaining 54.8 COCO mask AP. Experiments on LVIS, RefCOCO variants, and SA-V show consistent gains. A CLIP study also shows improved image-text retrieval.

Figures & tables

Appendix figures & tables8 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. CAR-SAM: Cross-Attention Reconstruction for Post-Training Quantization of the Segment Anything Model

    May 16, 2026Houji Wen, Jiangyong Yu, Jun Li +1Cross-Attention4-Bit Quantization

  2. ActiveSAM: Fast and Accurate Open-Vocabulary Semantic Segmentation with Frozen SAM 3

    Jun 15, 2026Tran Dinh Tien, Zhiqiang ShenEfficient VLM InferenceOpen-Vocabulary Semantic Segmentation