Measurement-informed Deep Learning for Real-time Monitoring of Hydrodynamic Drying Dynamics in Inkjet-printed Display Pixels via Low-resolution Imaging

Dataset OverviewThis repository contains the official image dataset used in the research article, "Measurement-informed Deep Learning for Real-time Monitoring of Hydrodynamic Drying Dynamics in Inkjet-printed Display Pixels via Low-resolution Imaging," recently accepted for publication in ACS Applied Materials & Interfaces. Purpose of the ProjectThe dataset was constructed to develop and validate a highly efficient Convolutional Neural Network (CNN) model capable of monitoring the hydrodynamic drying dynamics of inkjet-printed pixels in real-time. By utilizing low-resolution imaging, the dataset enables the model to accurately classify various drying stages (e.g., Classes A, B, C, α, and β) while ensuring computational efficiency for in-situ manufacturing process monitoring. Dataset StructureThe uploaded data is divided into two main compressed archives:* `Original_Image.zip`: Contains the primary dataset split randomly into training, validation, and test sets.* `Holdout_Image.zip`: Contains a strictly isolated hold-out dataset. This dataset was constructed from independent drying runs to robustly evaluate the generalizability and reliability of the trained model, completely isolating it from the original training distribution. Associated Source CodeThe deep learning training, computational efficiency evaluation, and Explainable AI (XAI) analysis codes that utilize this dataset are publicly available on GitHub. * GitHub Repository: [https://github.com/semidobee/] Usage & CitationIf you find this dataset useful for your research, please consider citing our corresponding ACS Applied Materials & Interfaces paper. (Complete citation details, including the DOI of the published article, will be updated here upon official online publication).