# EMA-SAM: Exponential Moving-average for SAM-based PTMC Segmentation

arXiv: https://arxiv.org/abs/2510.18213 · HTML: https://mlsanity.com/paper/2510.18213/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 34 references that are in the map.
- First posted on arXiv 21 Oct 2025.

## Builds on (most cited first, 23 of 23)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [Feature Pyramid Networks for Object Detection](https://mlsanity.com/paper/1612.03144/) (2016)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [UNet++: A Nested U-Net Architecture for Medical Image Segmentation](https://mlsanity.com/paper/1807.10165/) (2018)
- [Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks](https://mlsanity.com/paper/1710.11063/) (2017)
- [Medical SAM Adapter: Adapting Segment Anything Model for Medical Image Segmentation](https://mlsanity.com/paper/2304.12620/) (2023)
- [PraNet: Parallel Reverse Attention Network for Polyp Segmentation](https://mlsanity.com/paper/2006.11392/) (2020)
- [Road Extraction by Deep Residual U-Net](https://mlsanity.com/paper/1711.10684/) (2017)
- [Hiera: A Hierarchical Vision Transformer without the Bells-and-Whistles](https://mlsanity.com/paper/2306.00989/) (2023)
- [UniFormer: Unified Transformer for Efficient Spatiotemporal Representation Learning](https://mlsanity.com/paper/2201.04676/) (2022)
- [Medical SAM 2: Segment medical images as video via Segment Anything Model 2](https://mlsanity.com/paper/2408.00874/) (2024)
- [SAM2Long: Enhancing SAM 2 for Long Video Segmentation with a Training-Free Memory Tree](https://mlsanity.com/paper/2410.16268/) (2024)
- [Progressively Normalized Self-Attention Network for Video Polyp Segmentation](https://mlsanity.com/paper/2105.08468/) (2021)
- [SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation](https://mlsanity.com/paper/2408.08870/) (2024)
- [Vivim: a Video Vision Mamba for Medical Video Segmentation](https://mlsanity.com/paper/2401.14168/) (2024)
- [Surgical SAM 2: Real-time Segment Anything in Surgical Video by Efficient Frame Pruning](https://mlsanity.com/paper/2408.07931/) (2024)
- [Beyond Adapting SAM: Towards End-to-End Ultrasound Image Segmentation via Auto Prompting](https://mlsanity.com/paper/2309.06824/) (2023)
- [LDNet: Unified Listener Dependent Modeling in MOS Prediction for Synthetic Speech](https://mlsanity.com/paper/2110.09103/) (2021)
- [ACSNet: Action-Context Separation Network for Weakly Supervised Temporal Action Localization](https://mlsanity.com/paper/2103.15088/) (2021)
- [CC-SAM: SAM with Cross-feature Attention and Context for Ultrasound Image Segmentation](https://mlsanity.com/paper/2408.00181/) (2024)
- [Novel adaptation of video segmentation to 3D MRI: efficient zero-shot knee segmentation with SAM2](https://mlsanity.com/paper/2408.04762/) (2024)

## Abstract (arXiv)

Papillary thyroid microcarcinoma (PTMC) is increasingly managed with radio-frequency ablation (RFA), yet accurate lesion segmentation in ultrasound videos remains difficult due to low contrast, probe-induced motion, and heat-related artifacts. The recent Segment Anything Model 2 (SAM-2) generalizes well to static images, but its frame-independent design yields unstable predictions and temporal drift in interventional ultrasound. We introduce EMA-SAM, a lightweight extension of SAM-2 that incorporates a confidence-weighted exponential moving average pointer into the memory bank, providing a stable latent prototype of the tumour across frames. This design preserves temporal coherence through probe pressure and bubble occlusion while rapidly adapting once clear evidence reappears. On our curated PTMC-RFA dataset (124 minutes, 13 patients), EMA-SAM improves maxDice from 0.82 (SAM-2) to 0.86 and maxIoU from 0.72 to 0.76, while reducing false positives by 29%. On external benchmarks, including VTUS and colonoscopy video polyp datasets, EMA-SAM achieves consistent gains of 2--5 Dice points over SAM-2. Importantly, the EMA pointer adds textless0.1% FLOPs, preserving real-time throughput of ~30,FPS on a single A100 GPU. These results establish EMA-SAM as a robust and efficient framework for stable tumour tracking, bridging the gap between foundation models and the stringent demands of interventional ultrasound. Codes are available here hyperref[code https://github.com/mdialameh/EMA-SAM.