View-Consistent Hierarchical 3D Segmentation Using Ultrametric Feature Fields

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Haodi, Stearns, Colton, Harley, Adam W., Guibas, Leonidas J.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929425693540352
author He, Haodi
Stearns, Colton
Harley, Adam W.
Guibas, Leonidas J.
author_facet He, Haodi
Stearns, Colton
Harley, Adam W.
Guibas, Leonidas J.
contents Large-scale vision foundation models such as Segment Anything (SAM) demonstrate impressive performance in zero-shot image segmentation at multiple levels of granularity. However, these zero-shot predictions are rarely 3D-consistent. As the camera viewpoint changes in a scene, so do the segmentation predictions, as well as the characterizations of "coarse" or "fine" granularity. In this work, we address the challenging task of lifting multi-granular and view-inconsistent image segmentations into a hierarchical and 3D-consistent representation. We learn a novel feature field within a Neural Radiance Field (NeRF) representing a 3D scene, whose segmentation structure can be revealed at different scales by simply using different thresholds on feature distance. Our key idea is to learn an ultrametric feature space, which unlike a Euclidean space, exhibits transitivity in distance-based grouping, naturally leading to a hierarchical clustering. Put together, our method takes view-inconsistent multi-granularity 2D segmentations as input and produces a hierarchy of 3D-consistent segmentations as output. We evaluate our method and several baselines on synthetic datasets with multi-view images and multi-granular segmentation, showcasing improved accuracy and viewpoint-consistency. We additionally provide qualitative examples of our model's 3D hierarchical segmentations in real world scenes. The code and dataset are available at https://github.com/hardyho/ultrametric_feature_fields
format Preprint
id arxiv_https___arxiv_org_abs_2405_19678
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle View-Consistent Hierarchical 3D Segmentation Using Ultrametric Feature Fields
He, Haodi
Stearns, Colton
Harley, Adam W.
Guibas, Leonidas J.
Computer Vision and Pattern Recognition
Artificial Intelligence
Large-scale vision foundation models such as Segment Anything (SAM) demonstrate impressive performance in zero-shot image segmentation at multiple levels of granularity. However, these zero-shot predictions are rarely 3D-consistent. As the camera viewpoint changes in a scene, so do the segmentation predictions, as well as the characterizations of "coarse" or "fine" granularity. In this work, we address the challenging task of lifting multi-granular and view-inconsistent image segmentations into a hierarchical and 3D-consistent representation. We learn a novel feature field within a Neural Radiance Field (NeRF) representing a 3D scene, whose segmentation structure can be revealed at different scales by simply using different thresholds on feature distance. Our key idea is to learn an ultrametric feature space, which unlike a Euclidean space, exhibits transitivity in distance-based grouping, naturally leading to a hierarchical clustering. Put together, our method takes view-inconsistent multi-granularity 2D segmentations as input and produces a hierarchy of 3D-consistent segmentations as output. We evaluate our method and several baselines on synthetic datasets with multi-view images and multi-granular segmentation, showcasing improved accuracy and viewpoint-consistency. We additionally provide qualitative examples of our model's 3D hierarchical segmentations in real world scenes. The code and dataset are available at https://github.com/hardyho/ultrametric_feature_fields
title View-Consistent Hierarchical 3D Segmentation Using Ultrametric Feature Fields
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2405.19678