通过剔除简单样本，用数据瘦身提升PET/CT病灶分割精度。
Data Diet: Can Trimming PET/CT Datasets Enhance Lesion Segmentation?
- 基于模型损失筛选并移除最易分类样本，重构训练集。
- 使假阴性体积下降，初步测试集上骰子系数与假阳性均改善。
- 适合关注数据质量、模型鲁棒性的医学影像研究者。
本文针对autoPET3数据驱动赛道提出方法。传统观点认为数据量越大模型性能越好，但近期研究显示剔除部分样本可提升准确率。在autoPETIII数据集中，全量数据训练的模型对PSMA-PET产生大量假阳性。为此，我们依据模型损失值剔除训练集中最简单的样本，重新从头训练。该方法显著降低假阴性体积，并在初步测试集上优于基线模型，同时提升骰子系数。代码与预训练模型已开源至github.com/alexanderjaus/autopet3_datadiet。
原文摘要 · Abstract (English)
In this work, we describe our approach to compete in the autoPET3 datacentric track. While conventional wisdom suggests that larger datasets lead to better model performance, recent studies indicate that excluding certain training samples can enhance model accuracy. We find that in the autoPETIII dataset, a model that is trained on the entire dataset exhibits undesirable characteristics by producing a large number of false positives particularly for PSMA-PETs. We counteract this by removing the easiest samples from the training dataset as measured by the model loss before retraining from scratch. Using the proposed approach we manage to drive down the false negative volume and improve upon the baseline model in both false negative volume and dice score on the preliminary test set. Code and pre-trained models are available at github.com/alexanderjaus/autopet3_datadiet.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。