Deep Learning‐Enabled Imaging Flow Cytometry for High‐Speed Cryptosporidium and Giardia Detection

摘要: Imaging flow cytometry has become a popular technology for bioparticle images analysis because of its capability capturing thousands per second. Nevertheless, the vast number generated by imaging imposes great challenges data especially when species have similar morphologies. In this work, we report deep learning-enabled high-throughput system predicting Cryptosporidium and Giardia in drinking water. This combines an efficient artificial neural network called MCellNet, which achieves classification accuracy >99.6%. The can detect with sensitivity 97.37% specificity 99.95%. high-speed reaches 346 frames second, outperforming state-of-the-art learning algorithm MobileNetV2 speed (251 second) comparable accuracy. reported empowers rapid, accurate, high throughput detection clinical diagnostics, environmental monitoring other potential biosensing applications. article is protected copyright. All rights reserved.