Computer Vision-Based Artificial Intelligence-Mediated Encoding-Decoding for Multiplexed Microfluidic Digital Immunoassay

Digital immunoassays with multiplexed capacity, ultrahighsensitivity,and broad affordability are urgently required in clinical diagnosis,food safety, and environmental monitoring. In this work, a multidimensionaldigital immunoassay has been developed through microparticle-basedencoding and artificial intelligence-based decoding, enabling multiplexeddetection with high sensitivity and convenient operation. The informationencoded in the features of microspheres, including their size, number,and color, allows for the simultaneous identification and accuratequantification of multiple targets. Computer vision-based artificialintelligence can analyze the microscopy images for information decodingand output identification results visually. Moreover, the opticalmicroscopy imaging can be well integrated with the microfluidic platform,allowing for encoding-decoding through the computer vision-basedartificial intelligence. This microfluidic digital immunoassay cansimultaneously analyze multiple inflammatory markers and antibioticswithin 30 min with high sensitivity and a broad detection range frompg/mL to & mu;g/mL, which holds great promise as an intelligentbioassay for next-generation multiplexed biosensing.