小柯机器人

东南大学Hanchuan Peng、美国乔治梅森大学Giorgio A. Ascoli和澳大利亚新南威尔士大学Erik Meijering共同合作，近期取得重要工作进展。他们研究开发了BigNeuron工具，一种用于基准测试和预测光学显微镜数据集中神经元自动追踪算法性能的资源。相关研究成果2023年4月17日在线发表于《自然—方法学》杂志上。

BigNeuron是一个开放的社区台架测试平台，旨在为准确快速的自动神经元追踪设定开放标准。

研究人员收集了几个物种的不同图像集，这些图像集代表了许多对神经元追踪感兴趣的神经科学实验室获得的数据。研究人员报告了为可用成像数据集的子集生成的金标准手动注释，以及35种自动跟踪算法的量化跟踪质量。生成这样一个手工策划的多样化数据集的目标是推进跟踪算法的开发，并实现可推广的基准测试。与图像质量特征一起，研究人员将数据汇集在一个交互式web应用程序中，该应用程序使用户和开发人员能够执行主成分分析、t分布随机邻居嵌入、相关性和聚类、成像和跟踪数据的可视化，以及在用户定义的数据子集中对自动跟踪算法进行基准测试。图像质量指标解释了数据中的大部分差异，其次是与神经元大小相关的神经形态学特征。

研究人员观察到，不同的算法可以提供互补的信息来获得准确的结果，并开发了一种迭代组合并生成一致性重建的方法。所获得的一致性树提供了神经元结构基本事实的估计，其在噪声数据集中通常优于单个算法。然而，在特定的成像条件下，特定的算法可能优于共识树策略。

最后，为了帮助用户在没有手动注释进行比较的情况下预测最准确的自动跟踪结果，研究人员使用支持向量机回归来预测给定图像集和一组自动跟踪的重建质量。

附：英文原文

Title: BigNeuron: a resource to benchmark and predict performance of algorithms for automated tracing of neurons in light microscopy datasets

Author: Manubens-Gil, Linus, Zhou, Zhi, Chen, Hanbo, Ramanathan, Arvind, Liu, Xiaoxiao, Liu, Yufeng, Bria, Alessandro, Gillette, Todd, Ruan, Zongcai, Yang, Jian, Radojevi, Miroslav, Zhao, Ting, Cheng, Li, Qu, Lei, Liu, Siqi, Bouchard, Kristofer E., Gu, Lin, Cai, Weidong, Ji, Shuiwang, Roysam, Badrinath, Wang, Ching-Wei, Yu, Hongchuan, Sironi, Amos, Iascone, Daniel Maxim, Zhou, Jie, Bas, Erhan, Conde-Sousa, Eduardo, Aguiar, Paulo, Li, Xiang, Li, Yujie, Nanda, Sumit, Wang, Yuan, Muresan, Leila, Fua, Pascal, Ye, Bing, He, Hai-yan, Staiger, Jochen F., Peter, Manuel, Cox, Daniel N., Simonneau, Michel, Oberlaender, Marcel, Jefferis, Gregory, Ito, Kei, Gonzalez-Bellido, Paloma, Kim, Jinhyun, Rubel, Edwin, Cline, Hollis T., Zeng, Hongkui, Nern, Aljoscha, Chiang, Ann-Shyn, Yao, Jianhua, Roskams, Jane, Livesey, Rick, Stevens, Janine, Liu, Tianming, Dang, Chinh, Guo, Yike, Zhong, Ning, Tourassi, Georgia, Hill, Sean, Hawrylycz, Michael, Koch, Christof, Meijering, Erik, Ascoli, Giorgio A., Peng, Hanchuan

Issue&Volume: 2023-04-17

Abstract: BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. We observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.

DOI: 10.1038/s41592-023-01848-5

Source: https://www.nature.com/articles/s41592-023-01848-5

Nature Methods：《自然—方法学》，创刊于2004年。隶属于施普林格·自然出版集团，最新IF：47.99
官方网址：https://www.nature.com/nmeth/
投稿链接：https://mts-nmeth.nature.com/cgi-bin/main.plex