Data-driven constitutive modeling has the potential to redefine material mechanics, yet its progress is fundamentally bottlenecked by the difficulty of experimentally capturing rich, multiaxial stress–strain datasets. Conventional methods like Digital Image Correlation (DIC) provide full-field total strain distributions but lack...