Harnessing SWIR Hyperspectral Cameras for Contactless Cell Parameter Quantification

In the biomedical field, optical sensors play a crucial role in measuring cell parameters without labels. However, traditional optical probes require a lot of data to function correctly, which can be tedious to collect. Additionally, many professionals struggle to trust and troubleshoot complex models used in important tasks like biopharmaceutical manufacturing. Enter contactless short-wave infrared (SWIR) hyperspectral cameras. These cameras can capture thousands of absorption signals quickly, making it easier to remove noise and use simpler, interpretable models to measure metabolites.

Researchers tested a model called L-SLR with a SWIR HSI camera to quantify fructose, viable cell density (VCD), glucose, and lactate. Using only half of the data for training, the model performed well, with mean absolute errors (MAEs) and correlations (r²) close to expected values. For instance, it achieved an r² of 0. 88 and an MAE of 37 mg/dL for glucose. Even in complex media like Fetal Bovine Serum (FBS), the model could accurately measure fructose levels, both high (100-1000 mg/dL) and low (

Actions