Nature ( IF 48.5 ) Pub Date : 2025-06-04 , DOI: 10.1038/s41586-025-08991-3
Minhao Yang 1 , Rodney Douglas 2 , Chenghan Li 3 , Andreas Suess 4 , Hyunsurk Ryu 5 , Christoph Posch 6 , Shoushun Chen 7 , Raphael Berner 8 , Tetsuya Yagi 9 , Boyd Fowler 10 , Tobi Delbruck 2
![]() |
arising from: Z. Yang et al. Nature https://doi.org/10.1038/s41586-024-07358-4 (2024).
Biological eyes have many output cell types that provide different representations of a visual scene. Hybrid vision sensors with multi-pathway readout have emerged over the past decade for enhanced vision compared with conventional complementary metal–oxide–semiconductor (CMOS) image sensors. Yang et al.1 recently reported the frame-based vision sensor Tianmouc, which offers the combination of linear-dense colour images and linear-sparse temporal and spatial difference output. We appreciate the hard work that went into building the Tianmouc sensor and demonstrating its features. However, we call into question the novelty of the conceptual underpinnings and the claimed superior performance against the state of the art reported in the work by Yang et al.
中文翻译:

混合视觉传感器的动态范围和精度
来源 :Z. Yang 等人,Nature https://doi.org/10.1038/s41586-024-07358-4 (2024)。
生物眼具有许多输出单元类型,这些单元类型提供视觉场景的不同表示形式。与传统的互补金属氧化物半导体 (CMOS) 图像传感器相比,具有多路径读出功能的混合视觉传感器在过去十年中问世,用于增强视觉。Yang 等人 1 最近报道了基于帧的视觉传感器 Tianmouc,它提供了线性密集彩色图像和线性稀疏时空差异输出的组合。我们感谢为构建 Tianmouc 传感器并展示其功能而付出的辛勤工作。然而,我们质疑概念基础的新颖性,以及与 Yang 等人的工作中报告的最新技术相比所声称的优越性能。