# Rethinking Annotator Simulation: Realistic Evaluation of Whole-Body PET Lesion Interactive Segmentation Methods

arXiv: https://arxiv.org/abs/2404.01816 · HTML: https://mlsanity.com/paper/2404.01816/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 27 references that are in the map.
- First posted on arXiv 2 Apr 2024.

## Builds on (most cited first, 17 of 17)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning](https://mlsanity.com/paper/1506.02142/) (2015)
- [The Medical Segmentation Decathlon](https://mlsanity.com/paper/2106.05735/) (2021)
- [TotalSegmentator: robust segmentation of 104 anatomical structures in CT images](https://mlsanity.com/paper/2208.05868/) (2022)
- [AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation](https://mlsanity.com/paper/2206.08023/) (2022)
- [Interactive Medical Image Segmentation using Deep Learning with Image-specific Fine-tuning](https://mlsanity.com/paper/1710.04043/) (2017)
- [DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation](https://mlsanity.com/paper/1707.00652/) (2017)
- [Large-scale interactive object segmentation with human annotators](https://mlsanity.com/paper/1903.10830/) (2019)
- [ISLES 2022: A multi-center magnetic resonance imaging stroke lesion segmentation dataset](https://mlsanity.com/paper/2206.06694/) (2022)
- [MIDeepSeg: Minimally Interactive Segmentation of Unseen Objects from Medical Images Using Deep Learning](https://mlsanity.com/paper/2104.12166/) (2021)
- [MONAI Label: A framework for AI-assisted Interactive Labeling of 3D Medical Images](https://mlsanity.com/paper/2203.12362/) (2022)
- [iSegFormer: Interactive Segmentation via Transformers with Application to 3D Knee MR Images](https://mlsanity.com/paper/2112.11325/) (2021)
- [Deep Interactive Segmentation of Medical Images: A Systematic Review and Taxonomy](https://mlsanity.com/paper/2311.13964/) (2023)
- [Multimodal Interactive Lung Lesion Segmentation: A Framework for Annotating PET/CT Images based on Physiological and Anatomical Cues](https://mlsanity.com/paper/2301.09914/) (2023)
- [Guiding the Guidance: A Comparative Analysis of User Guidance Signals for Interactive Segmentation of Volumetric Images](https://mlsanity.com/paper/2303.06942/) (2023)
- [Adaptive Multi-scale Online Likelihood Network for AI-assisted Interactive Segmentation](https://mlsanity.com/paper/2303.13696/) (2023)
- [Sliding Window FastEdit: A Framework for Lesion Annotation in Whole-body PET Images](https://mlsanity.com/paper/2311.14482/) (2023)

## Built on by (most cited first, 2 of 2)

- [From Static to Interactive: Adapting Visual in-Context Learners for User-Driven Tasks](https://mlsanity.com/paper/2604.06748/) (2026)
- [IMPACT-Scribe: Interactive Temporal Action Segmentation with Boundary Scribbles and Query Planning](https://mlsanity.com/paper/2605.01668/) (2026)

## Abstract (arXiv)

Interactive segmentation plays a crucial role in accelerating the annotation, particularly in domains requiring specialized expertise such as nuclear medicine. For example, annotating lesions in whole-body Positron Emission Tomography (PET) images can require over an hour per volume. While previous works evaluate interactive segmentation models through either real user studies or simulated annotators, both approaches present challenges. Real user studies are expensive and often limited in scale, while simulated annotators, also known as robot users, tend to overestimate model performance due to their idealized nature. To address these limitations, we introduce four evaluation metrics that quantify the user shift between real and simulated annotators. In an initial user study involving four annotators, we assess existing robot users using our proposed metrics and find that robot users significantly deviate in performance and annotation behavior compared to real annotators. Based on these findings, we propose a more realistic robot user that reduces the user shift by incorporating human factors such as click variation and inter-annotator disagreement. We validate our robot user in a second user study, involving four other annotators, and show it consistently reduces the simulated-to-real user shift compared to traditional robot users. By employing our robot user, we can conduct more large-scale and cost-efficient evaluations of interactive segmentation models, while preserving the fidelity of real user studies. Our implementation is based on MONAI Label and will be made publicly available.