FrozenRecon: Pose-free 3D Scene Reconstruction with Frozen Depth Models

TL;DR

This paper introduces a test-time optimization method that leverages pre-trained affine-invariant depth models to achieve robust, pose-free 3D scene reconstruction across diverse datasets, with minimal parameter tuning per video.

Contribution

It proposes a novel approach that freezes large-scale depth models and optimizes only a few parameters for consistent 3D reconstruction without requiring camera poses.

Findings

Achieves state-of-the-art results on five zero-shot datasets.

Ensures inter-frame depth consistency in challenging scenes.

Operates with only dozens of parameters per frame.

Abstract

3D scene reconstruction is a long-standing vision task. Existing approaches can be categorized into geometry-based and learning-based methods. The former leverages multi-view geometry but can face catastrophic failures due to the reliance on accurate pixel correspondence across views. The latter was proffered to mitigate these issues by learning 2D or 3D representation directly. However, without a large-scale video or 3D training data, it can hardly generalize to diverse real-world scenarios due to the presence of tens of millions or even billions of optimization parameters in the deep network. Recently, robust monocular depth estimation models trained with large-scale datasets have been proven to possess weak 3D geometry prior, but they are insufficient for reconstruction due to the unknown camera parameters, the affine-invariant property, and inter-frame inconsistency. Here, we propose a novel test-time optimization approach that can transfer the robustness of affine-invariant depth models such as LeReS to challenging diverse scenes while ensuring inter-frame consistency, with only dozens of parameters to optimize per video frame. Specifically, our approach involves freezing the pre-trained affine-invariant depth model's depth predictions, rectifying them by optimizing the unknown scale-shift values with a geometric consistency alignment module, and employing the resulting scale-consistent depth maps to robustly obtain camera poses and achieve dense scene reconstruction, even in low-texture regions. Experiments show that our method achieves state-of-the-art cross-dataset reconstruction on five zero-shot testing datasets.