Preprint
Aug 2026
Efficient and High-Quality Depth Estimation via Pixel-Space Diffusion with Linear Attention
Extensive evaluations on multiple benchmarks demonstrate that Lapis consistently achieves state-of-the-art (SOTA) accuracy and boundary sharpness across various resolutions, reducing inference latency by up to 7.6 times compared to previous SOTA generative models.
Bing-De Liu, Wu Ran, Jinglei Zhang et al.
· 1 citation