Prompt-based Generation of 3D Textured Human with Motion Rigging

Recent advances in domains such as High Performance Computing (HPC) and proliferation of Machine Learning have led to great strides in the domain of 3D model generation. Several technologies, such as the Metaverse, AR/VR applications and game development have benefited tremendously from this phenomenon as it makes the tedious and time-consuming process of 3D modeling convenient. In real-world applications such as gaming, virtual reality, and augmented reality, 3D models with predefined articulated structures are essential for immersive experiences and lifelike animations. The solution proposed in this paper aims to expedite the arduous process of manually annotating articulation points within 3D models, a prevalent necessity with current models. By integrating an automated articulation definition mechanism into our model, we streamline the animation workflow and enhance user efficiency. With a CLIP Score of 0.548, the 3D model itself is sufficiently high quality enough to serve this purpose. This functionality enables expedited animation creation and fosters smoother workflow by providing predefined articulation points within the 3D models.