Machine learning-enabled flexible luminescent sensor for non-destructive mapping antibiotics distribution on seafood

文献类型： 外文期刊

第一作者： Wu, Lidong

作者： Wu, Lidong;Li, Yuanxin;Qin, Haiyang;Zhao, Jinxue;Zhai, Xuejing;Li, Peiyi;Wu, Lidong;Li, Yuanxin;Qin, Haiyang;Zhao, Jinxue;Zhai, Xuejing;Li, Peiyi;Wu, Lidong;Li, Yuanxin;Zhai, Xuejing;Li, Peiyi;Li, Zhibo;Wu, Lidong;Qin, Haiyang;Zhao, Jinxue;Li, Peiyi;Xiang, Xueping

作者机构：

关键词： Luminescent hydrogel sensors; Machine learning; Non-destructive identification; Concentration-distribution patterns; Tetracycline

期刊名称：CHEMICAL ENGINEERING JOURNAL （ 影响因子：13.2； 五年影响因子：13.5 ）

ISSN： 1385-8947

年卷期： 2025 年 510 卷

页码：

收录情况： SCI

摘要： Fast, simultaneous, and non-destructive identification of antibiotic distribution on food is critical to public health. Here We present a novel methodology for mapping antibiotics using luminescent hydrogel sensors (LHS) enhanced by machine learning. The LHS comprises a hydrogel substrate with luminescent europium ions (Eu3+), undergoing chromatic alterations upon exposure to tetracycline (TC). These luminescent spectral variations are captured using a smartphone and used to train convolutional neural networks (CNNs), resulting in high accuracy (exceeding 92 %) for identifying and quantifying concentration-distribution patterns. The LHS displays a linear detection range of TC (5 similar to 100 mu g L-1) with a detection limit of 0.58 mu g L-1. The trained LHS-CNN system can simultaneously delineate the distribution of TC on cross-sections of animal-derived foods or moistened animal skin, representing a realistic and complex environmental setting. This method has the potential to advance non-destructive detection and identification of molecules on food substrates, including those with curved surfaces, without the need for enrichment or additional sample preparation.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序