Key points are not available for this paper at this time.
Colorimetric assays offer a low-cost, accessible means of diagnostic testing but often suffer from subjective interpretation and variability caused by inconsistent imaging conditions. To address these challenges, we present HueTools, a comprehensive image processing software that enables quantitative development of paper-based colorimetric assays using smartphone imaging. HueTools integrates a mobile app, and a web platform to standardize image capture, apply precise color correction, and predictive modeling of analyte concentration using interpretable machine learning. The system supports a complete workflow: from image acquisition and color calibration to region-of-interest (ROI) selection, signal extraction, and statistical analysis. The system enables perceptually grounded color analysis using the CIELAB space and provides quantitative metrics, including LoB, LoD, and LoQ via interpretable ensemble-based customization models. HueTools was validated using a lateral flow dipstick luteinizing hormone (LH) and vertical flow alanine transaminase (ALT) assays. The results demonstrate HueTools' ability to reduce human error, improve assay reproducibility, and provide feedback for optimizing assay design. It also supports seamless transitions between field testing and lab analysis, allowing researchers to capture images on-site and perform in-depth analysis remotely. HueTools offers a hardware-independent, cloud-based solution for assay developers, streamlining workflows while minimizing costs associated with dedicated readers. Its accessibility, automation, and cross-platform compatibility make it well-suited for research and development of colorimetric point-of-care diagnostics, especially in resource-limited settings.
Sharma et al. (Thu,) studied this question.