Efficient and High-Quality Depth Estimation via Pixel-Space Diffusion with Linear Attention

arXiv:2608.30129v1 Announce Type: new Abstract: This work presents extbf{Lapis}, a extbf{l}inear-extbf{a}ttention-based extbf{pi}xel-extbf{s}pace generative framework that achieves efficient and high-fidelity depth estimation with one-step diffusion. While generative frameworks have significantly advanced monocular depth estimation with superior detail fidelity, the O(N^2) complexity of standard attention and the multi-step denoising process introduce prohibitive computational costs when scaling them to high-resolution image applications. Although linear attention and one-step prediction are intuitively viable, directly applying them leads to poor structural consistency, detail loss, and noise. Lapis rectifies these limitations through a coarse-to-fine hierarchy. Specifically, a Patch-level Consistency Module restores structural coherence by integrating semantic and spatial priors. Subsequently, a Pixel-level Refinement Module recovers sharp geometric boundaries via skip-connection-based pixel correspondence. Furthermore, to mitigate sampling noise inherent in one-step diffusion, we leverage the manifold assumption and adopt a direct mathbf{x}-prediction strategy to target the clean data manifold. 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 7.6imes at 1080P and 10.9imes at 1440P resolution compared to previous SOTA generative models.

Source: arXiv cs.CV | 2026-09-01

Related