Cortex-Synth: Differentiable Topology-Aware 3D Skeleton Synthesis with Hierarchical Graph Attention
This work addresses the challenge of accurate 3D skeleton reconstruction for applications like robotic manipulation, medical imaging, and character rigging, representing a strong specific gain rather than a broad paradigm shift.
The paper tackled the problem of synthesizing 3D skeleton geometry and topology from single 2D images, achieving state-of-the-art results with an 18.7% improvement in MPJPE and a 27.3% improvement in Graph Edit Distance on ShapeNet, while reducing topological errors by 42%.
We present Cortex Synth, a novel end-to-end differentiable framework for joint 3D skeleton geometry and topology synthesis from single 2D images. Our architecture introduces three key innovations: (1) A hierarchical graph attention mechanism with multi-scale skeletal refinement, (2) Differentiable spectral topology optimization via Laplacian eigen decomposition, and (3) Adversarial geometric consistency training for pose structure alignment. The framework integrates four synergistic modules: a pseudo 3D point cloud generator, an enhanced PointNet encoder, a skeleton coordinate decoder, and a novel Differentiable Graph Construction Network (DGCN). Our experiments demonstrate state-of-the-art results with 18.7 percent improvement in MPJPE and 27.3 percent in Graph Edit Distance on ShapeNet, while reducing topological errors by 42 percent compared to previous approaches. The model's end-to-end differentiability enables applications in robotic manipulation, medical imaging, and automated character rigging.