Dataset: Image-based deep-learning analysis of laminar flame-speed landscapes in C1-C3 hydrocarbon/air conical flames

This dataset contains all raw images (.tiff), metadata (.json), scripts used to train the models (CNN only, hybrid, generalization) for the studies presented in the paper, and all results as .xlsx file. test_predict.xlsx has the summarized metadata for all conditions, which is used to get all the relevant parameters for training (ground truth flame speeds from 1D simulations, unburned gas temperature, equivalence ratio, corrected gas exit velocity, .... train_main_models.py trains the image only CNN and the hybrid (CNN+MLP) model train_generalization.py trains the models which are used for the study summarized in Table 2 in the main manuscript.