6 papers
Diffusion Image Generation with Explicit Modeling of Data Manifold Geometry
Duoduo Xue, Zhiyu Zhu, Junhui Hou
Image generative models aim to sample data points from the underlying data manifold, a task that requires learning and decoding a dense, low-dimensional, and compact parameterizati…
AgentCompile: An LLM-Guided Compiler for Direct CUDA Inference
Xuanzhe Li, Ziyan Weng, Zhiyu Zhu +1
Transformer inference increasingly relies on specialized compiler and runtime support, while recent LLMs can generate nontrivial CUDA kernels. However, unconstrained generation gua…
Energy-oriented Diffusion Bridge for Image Restoration with Foundational Diffusion Models
Jinhui Hou, Zhiyu Zhu, Junhui Hou
Diffusion bridge models have shown great promise in image restoration by explicitly connecting clean and degraded image distributions. However, they often rely on complex and high-…
Scaling Dense Event-Stream Pretraining from Visual Foundation Models
Zhiwen Chen, Junhui Hou, Zhiyu Zhu +2
Learning versatile, fine-grained representations from irregular event streams is pivotal yet nontrivial, primarily due to the heavy annotation that hinders scalability in dataset s…
Spatially-guided Temporal Aggregation for Robust Event-RGB Optical Flow Estimation
Qianang Zhou, Junhui Hou, Meiyi Yang +3
Current optical flow methods exploit the stable appearance of frame (or RGB) data to establish robust correspondences across time. Event cameras, on the other hand, provide high-te…
ResFlow: Fine-tuning Residual Optical Flow for Event-based High Temporal Resolution Motion Estimation
Qianang Zhou, Zhiyu Zhu, Junhui Hou +3
Event cameras hold significant promise for high-temporal-resolution (HTR) motion estimation. However, estimating event-based HTR optical flow faces two key challenges: the absence…