发表机构
Institute of Computer Science, Warsaw University of Technology; Nencki Institute of Experimental Biology of the Polish Academy of Sciences; Medical University of Lublin; Medical University of Warsaw(华沙理工大学计算机科学研究所; 波兰科学院恩基实验生物学研究所; 卢布林医科大学; 华沙医科大学)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
研究提出三种用于斑马鱼表型分类的分层集成设置,第一阶段用四类分类器初分,“其他”类再经不同集成设计处理,用三种骨干架构比较,结果显示ConvNeXt性能最佳,专用分层集成有效,为表型识别提供新方法。
AI 中文摘要
我们提出并评估了三种用于从胚胎图像中对斑马鱼表型进行分类的分层集成设置。在所有设置中,第一阶段使用单个四类分类器将图像分配到以下排他性表型之一:正常、有绒毛膜、死亡或其他。分类为“其他”的图像然后在第二阶段进行处理,各设置中的集成设计有所不同:单个多标签分类器、两个专用多标签分类器或二元分类器的集成。我们使用三种骨干架构(ResNet18、ViT 和 ConvNeXt)比较了这些设置。总体而言,ConvNeXt 在所有设置中实现了最佳性能,而设置 2 中的专用分层集成在 F1 分数方面提供了最佳平衡。结果表明,所提出的专用分层集成对于斑马鱼表型识别是有效的,并表明 ConvNeXt 是特别有用的骨干模型。
英文摘要
We propose and evaluate three hierarchical ensemble setups for zebrafish phenotype classification from embryo images. In all setups, stage 1 uses a single four-class classifier to assign images to one of the exclusive phenotypes: Normal, Chorion, Dead, or Other. Images classified as Other are then processed in stage 2, where the ensemble design differs across setups: a single multi-label classifier, two specialized multi-label classifiers, or an ensemble of binary classifiers. We compare these setups using three backbone architectures: ResNet18, ViT, and ConvNeXt. Overall, ConvNeXt achieves the best performance across setups, while the specialized hierarchical ensemble in setup 2 provides the best balance in terms of F1-score. The results show that the proposed specialised hierarchical ensembles are effective for zebrafish phenotype recognition, and suggest that ConvNeXt is particularly useful backbone model.
CommentsAccepted to KES 2026 conference