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

This repository contains the official dataset and source code for the study titled "Physics-augmented Deep Learning for Real-time Monitoring of Hydrodynamic Drying Dynamics in Inkjet-printed Display Pixels via Low-resolution Imaging". This work is currently submitted to Advanced Science for publication. The provided resources are designed to ensure the full reproducibility of the experimental results, including the training of a lightweight Convolutional Neural Network (CNN) and the extraction of Explainable AI (XAI) insights. The framework achieves a 99.4% classification accuracy for the drying stages of ZnO nanoparticle inks on 200-PPI substrates while maintaining a 16.4-fold reduction in computational cost. The repository includes the following items: Image_data.zip: A comprehensive dataset of low-resolution optical images capturing the hydrodynamic evolution of ZnO nanoparticle inks (Stage A, B, C, alpha, and beta). Drying_Classifier_CNN.py: Source code for model training, validation, and performance evaluation. CLassifier_Parameters_Analysis.py: Script for calculating theoretical FLOPs, trainable parameters, and real-time inference latency. Classifier_XAI_Analysis.py: Tools for generating Input Gradient Saliency maps and Grad-CAM (Gradient-weighted Class Activation Mapping) overlays. requirements.txt: List of Python dependencies required for the environment setup. README.md: Detailed instructions for environment configuration and code execution. Access Condition:Access to the files is currently restricted for the peer-review process. The repository will be made publicly available under an open-access license upon the official publication of the manuscript.