3D Representations

Project Overview

The world is three-dimensional. We are broadly interested in processing and synthesis of 3D data. One line of work concerns 3D reconstruction, with the aim of producing photorealistic models of objects and environments. Such photorealistic models are used in virtual reality, special effects production, and computer games, and can also be used in training visual perception and sensorimotor control systems. Another line of work concerns 3D data processing. We are interested both in classic optimization techniques and in novel deep network architectures that operate on three-dimensional representations.

Publications

ASH: A Modern Framework for Parallel Spatial Hashing in 3D Perception

Shape from Polarization for Complex Scenes in the Wild

Geometry Processing with Neural Fields

Point Transformer

Adaptive Surface Reconstruction with Multiscale Convolutional Kernels

Self-supervised Geometric Perception

Deep Global Registration

High-dimensional Convolutional Networks for Geometric Pattern Recognition

On Joint Estimation of Pose, Geometry and svBRDF from a Handheld Scanner

Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-Shot Cross-Dataset Transfer

Fully Convolutional Geometric Features

Consensus Maximization Tree Search Revisited

What Do Single-view 3D Reconstruction Networks Learn?

Connecting the Dots: Learning Representations for Active Monocular Depth Estimation

Tangent Convolutions for Dense Prediction in 3D

Open3D: A Modern Library for 3D Data Processing

Learning Compact Geometric Features

Colored Point Cloud Registration Revisited

Tanks and Temples: Benchmarking Large-Scale Scene Reconstruction

Fast Global Registration

A Large Dataset of Object Scans

Robust Reconstruction of Indoor Scenes

Single-View Reconstruction via Joint Analysis of Image and Shape Collections

Robust Nonrigid Registration by Convex Optimization

Depth Camera Tracking with Contour Cues

Color Map Optimization for 3D Reconstruction with Consumer Depth Cameras

Fast MRF Optimization with Application to Depth Reconstruction

Simultaneous Localization and Calibration: Self-Calibration of Consumer Depth Cameras

Elastic Fragments for Dense Scene Reconstruction

A Simple Model for Intrinsic Image Decomposition with Depth Cues

Dense Scene Reconstruction with Points of Interest

Joint Shape Segmentation with Linear Programming