WonderJourney is capable of generating a series of continuous and diverse 3D scenes from a given starting point. Experience a smooth visual journey, where each scene is coherently connected to the previous one.
By leveraging a Large Language Model (LLM) for scene description generation, WonderJourney can comprehend and transform complex textual descriptions, creating visual scenes that match these descriptions.
A Visual Language Model (VLM) is used to ensure that the generated scenes not only conform to the descriptions but are also coherent and of high quality in visual terms. This enhances the appeal and realism of the generated content.
According to user preference assessments, WonderJourney excels in diversity, visual quality, scene complexity, and overall appeal, surpassing other existing methods. It has broad application potential in fields.
Begin a 3D journey with WonderJourney, starting from any location, described via text or image. It creates a 'wonderjourney'—a series of diverse, yet coherently linked 3D scenes along a camera path, rendered using a dynamic back-and-forth trajectory. This innovative approach transforms 3D scene generation, enhancing digital storytelling and virtual experiences.