AI-augmented ultrasound analysis of noninvasive quantification of hydrogels concentration for bioprinting

Abstract Hydrogels, possessing biocompatibility and flexibility, are widely used across biomedical and industrial domains, with their concentration serving as a critical determinant of their physicochemical properties. However, conventional methods for concentration assessment exhibit significant limitations; invasive techniques damage the original state of the sample, while existing non-invasive approaches often lack precision at extreme concentration levels. To address these challenges, this study introduces a novel, highly accurate, non-invasive ultrasound-based methodology for hydrogel concentration analysis. A single-element ultrasound transducer was used to collect concentration data while preserving sample integrity. This approach mitigates the accuracy variation observed in existing technologies, enabling precise classification across all concentration levels. In particular, complex ultrasound signal pattern analysis was conducted using a convolutional neural network-based machine learning framework, achieving concentration classification with an accuracy exceeding 99%. Through highly accurate and non-destructive concentration classification, the proposed method holds substantial potential as a core technology for improving the quality control of hydrogel-based constructs.