FreeView improves 3D scene reconstruction from messy photo collections
FreeView: Robust 3D Reconstruction from Uncurated Views
Computer Vision and Pattern Recognition
Summary
Creating 3D models from random photos can be hard because some images may not fit well or show unrelated views. The authors present FreeView, a tool that figures out which photos belong together without needing a carefully chosen example image. It does this by finding a good representative photo and filtering out unhelpful ones using clues from how pictures look and their geometry. This method helps build accurate 3D scenes even when the photo collection is mixed up.
What this means in practice
- •For photogrammetry teams: Automatically filter out irrelevant photos in large photo sets to produce better 3D reconstructions without manual reference selection.
- •For augmented reality developers: Integrate robust view filtering to ensure reliable 3D scene building from user-provided photo collections with varying viewpoints.
Authors
Junho Kim, Hyunyoung Choi, Seongwon Lee
Abstract
Reconstructing a coherent 3D scene from uncurated image collections remains challenging due to the presence of irrelevant or non-overlapping views. Recent work has shown that feed-forward reconstruction models contain internal view-consistency signals that support distractor rejection. However, reference-conditioned filtering relies on a curated clean reference, which is not guaranteed in practical image collections. To address this limitation, we introduce FreeView, a training-free framework that jointly selects a filtering reference and rejects distractor views from an uncurated collection. FreeView leverages collection-level consensus to identify a representative view that serves as a reliable reference for filtering. It then adapts view rejection to each input collection by combining complementary visual and geometric evidence. This approach preserves useful views under substantial appearance and viewpoint variations while rejecting distractors. FreeView integrates with both VGGT and pi^3, removing the need for a manually curated reference or manual calibration of view rejection. Extensive experiments across multiple datasets demonstrate high clean-view retention and distractor rejection rates, leading to reliable 3D reconstruction from uncurated views.