School of Materials Science and Innovation, Faculty of Science, Mahidol University, Nakhon Pathom, 73170, Thailand.
Department of Science and Technology, Nakhon Ratchasima Rajabhat University, Nakhon Ratchasima, 30000, Thailand.
Sci Rep. 2024 Sep 3;14(1):20498. doi: 10.1038/s41598-024-71238-0.
Rapid and accurate bacterial identification is essential for timely treatment of infections like sepsis. While traditional methods are reliable, they lack speed, and advanced molecular techniques often suffer from cost and complexity. The bacterial detection technology based on optical scattering system offers a rapid, label-free alternative but traditionally relies on complex lasers and analysis. Our enhanced approach utilizes RGB light emitting diodes (LEDs) as the light source. Three diffraction images of bacterial colonies from different LED colors are separately captured by a USB camera and combined using an image registration algorithm to enhance image sharpness. Our approach utilizes an object detection model, i.e., YOLOv8, for analysis achieving high-accuracy differentiation between bacterial strains. We demonstrate the effectiveness of this approach, achieving an average accuracy of 97% (mAP50 of 0.97), including accurate discrimination of closely related strains and the significant pathogen Staphylococcus aureus MRSA 1320. Our enhancement offers advantages in affordability, usability, and seamless integration into existing workflows, providing an alternative for rapid bacterial identification.
快速、准确的细菌鉴定对于败血症等感染的及时治疗至关重要。虽然传统方法可靠，但速度较慢，而先进的分子技术往往受到成本和复杂性的限制。基于光学散射系统的细菌检测技术提供了一种快速、无标记的替代方法，但传统上依赖于复杂的激光和分析。我们的增强方法利用 RGB 发光二极管（LED）作为光源。通过 USB 摄像头分别捕获来自不同 LED 颜色的细菌菌落的三个衍射图像，并使用图像配准算法将其组合以增强图像锐度。我们的方法利用目标检测模型，即 YOLOv8，进行分析，实现了对细菌菌株的高精度区分。我们证明了这种方法的有效性，实现了平均准确率 97%（mAP50 为 0.97），包括对密切相关菌株和重要病原体金黄色葡萄球菌 MRSA 1320 的准确区分。我们的增强方法在价格、可用性和与现有工作流程的无缝集成方面具有优势，为快速细菌鉴定提供了一种替代方案。
