Bayesian model for semi-automated zooplankton classification with predictive confidence and rapid category aggregation.

[謝志豪]

台灣大學海洋研究所副教授謝志豪以及其博士後研究員和學生發展出半自動浮游動物的分類系統,此項結果被發表在2011年11月的Marine Ecology Progress Series,這種方法已成功應用於中國東海的生態系統。這種半自動化系統可以分辨出較多的罕見種類,因此可提高浮游動物種數估計的準確度。此方法提升了浮游動物分類的效率,未來將能夠廣泛應用於環境監測以及生態研究。

圖一、ZooScan系統以及浮游動物影像
圖二、半自動化浮游動物分類系統所依據的統計分析方法

Reference: Ye, L., C. Y. Chang, and C. H. Hsieh (2011) Bayesian model for semi-automated zooplankton classification with predictive confidence and rapid category aggregation. Marine Ecology Progress Series. 441: 185-196[Chih-hao Hsieh]

Associate professor Chih-hao Hsieh from the Institute of Oceanography, NTU, leads his postdoc and student to develop an efficient semi-automatic zooplankton classification system. The results are published in the November issue of Marine Ecology Progress Series in 2012. This approach was successfully applied in the East China Sea ecosystem. In particular, the semi-automated approach achieves significant improvement in recognition accuracy for rare categories, thereby improving the estimates of taxa richness. This approach can make up for deficiencies of current automated zooplankton classifiers and facilitates an efficient semi-automated zooplankton classification, which may have a broad application in environmental monitoring and ecological research.

Figure 1. The ZooScan system and classified zooplankton images.
Figure 2. The statistical procedure to estimate classification confidence in order to achieve the semi-automatic classification.

Reference: Ye, L., C. Y. Chang, and C. H. Hsieh (2011) Bayesian model for semi-automated zooplankton classification with predictive confidence and rapid category aggregation. Marine Ecology Progress Series. 441: 185-196