Description:

This Master Prompt integrates digital image processing with machine learning, deep learning and computer vision for classification, detection, segmentation, reconstruction, quantitative prediction and research interpretation. It covers research-problem formulation, dataset construction, annotation, preprocessing, augmentation, class imbalance, feature engineering, model selection, training, hyperparameter optimization, validation, transfer learning, generalization, explainability, uncertainty and reproducible implementation. Particular emphasis is placed on establishing whether images serve as direct model inputs, sources of extracted predictors, sources of reference measurements or components of a hybrid analytical workflow.