A master's student whose research interests focus on MultiAgent, AIGC, and Avatar.
我是 YJY，目前在深圳。研究兴趣主要为 MultiAgent、AIGC、Avatar，研究方向涉及 新视角生成、时序影像重建、深度估计、超写实数字人 与 深度强化学习加速求解。
- Master's student focused on research-oriented and engineering-oriented AI systems
- Interested in MultiAgent systems, AIGC pipelines, avatar-related generation, and geometry-aware learning
- Building practical pipelines that connect model design, data processing, training, inference, and visualization
- MultiAgent
- AIGC
- Avatar
- Novel View Synthesis
- Temporal Image Reconstruction
- Depth Estimation
- Photorealistic Digital Humans
- Deep Reinforcement Learning Acceleration
- MALA
- Temporal-spatial collaborative reconstruction framework for remote sensing image completion.
- VDA-Absolute-Depth-Distillation
- Absolute depth distillation pipeline for robust monocular depth estimation.
- FlashAvatar-DepthFusion
- Monocular-depth-regularized FlashAvatar for head avatar reconstruction.
- depth_warp_vs
- Depth-guided warp-and-refine pipeline for monocular novel-view synthesis and naked-eye 3D rendering.
- VMAE
- Video masked autoencoder framework for interpolation and reconstruction.
- MultiAgent_workflow
- Experiments and tools for multi-agent workflow design and automation.
MultiAgent AIGC Avatar Novel View Synthesis Temporal Reconstruction Depth Estimation Digital Human Deep Reinforcement Learning
- MultiAgent workflow design and agent collaboration
- Novel view synthesis and geometry-aware generation
- Temporal image reconstruction under complex missing patterns
- Monocular depth estimation and depth distillation
- Photorealistic digital human and avatar reconstruction
- Accelerated solving with deep reinforcement learning
- Homepage: github.com/jinyuanyu
- Featured repositories are continuously updated here as the projects evolve
欢迎交流与合作。