论文标题
EPILLID数据集：用于药丸识别的低发细粒基准
ePillID Dataset: A Low-Shot Fine-Grained Benchmark for Pill Identification
论文作者
论文摘要
对于患者和医疗专业人员来说，识别处方药是一项常见的任务；但是，这是一项容易出错的任务，因为许多药丸具有相似的外观（例如白色圆形药丸），从而增加了药物错误的风险。在本文中，我们介绍了Epillid，这是药丸图像识别上最大的公共基准，由代表9804外观类别的13k图像组成（4902次药丸类型的两侧）。对于大多数外观类别，只有一个参考图像，使其成为一个具有挑战性的低射击识别设置。我们在基准上介绍了各种基线模型的实验设置和评估结果。使用双线性特征的多头度量学习方法的最佳基线表现出色。但是，我们的错误分析表明他们仍然无法区分特别令人困惑的类。代码和数据可在https://github.com/usuyama/epillid-benchmark上获得。
Identifying prescription medications is a frequent task for patients and medical professionals; however, this is an error-prone task as many pills have similar appearances (e.g. white round pills), which increases the risk of medication errors. In this paper, we introduce ePillID, the largest public benchmark on pill image recognition, composed of 13k images representing 9804 appearance classes (two sides for 4902 pill types). For most of the appearance classes, there exists only one reference image, making it a challenging low-shot recognition setting. We present our experimental setup and evaluation results of various baseline models on the benchmark. The best baseline using a multi-head metric-learning approach with bilinear features performed remarkably well; however, our error analysis suggests that they still fail to distinguish particularly confusing classes. The code and data are available at https://github.com/usuyama/ePillID-benchmark.
