Abstract:
We present the pioneering Large Visual Motion Model~(LVMM), meticulously engineered to analyze the intrinsic dynamics encapsulated within real-world imagery. Our model, fortified with a wealth of prior knowledge extracted from billions of image pairs, demonstrates promising results in predicting a diverse spectrum of scene dynamics. As a result, it can infuse any generic image with authentic dynamic effects, enhancing its visual allure. For a more comprehensive view of our results, please visit our project page: \url{https://anonymous.4open.science/r/AGI-LVMM-7826}.
Chat is not available.