Image-based modeling is a methodological approach or research domain centered on the construction of models or the extraction of actionable information derived directly and primarily from image data. This field investigates techniques for leveraging visual input to create representations (e.g., 3D structures, learned features) or predictive systems, encompassing areas such as geometric reconstruction from multiple views, self-supervised representation learning, and visual feature-based predictive analytics, enabling the digital representation, analysis, and prediction of real-world entities or processes based predominantly on visual sources.