Search Results for author: Andreas Geiger

Found 81 papers, 42 papers with code

Panoptic NeRF: 3D-to-2D Label Transfer for Panoptic Urban Scene Segmentation

no code implementations29 Mar 2022 Xiao Fu, Shangzhan Zhang, Tianrun Chen, Yichong Lu, Lanyun Zhu, Xiaowei Zhou, Andreas Geiger, Yiyi Liao

In this work, we present a novel 3D-to-2D label transfer method, Panoptic NeRF, which aims for obtaining per-pixel 2D semantic and instance labels from easy-to-obtain coarse 3D bounding primitives.

Instance Segmentation Scene Segmentation

TensoRF: Tensorial Radiance Fields

1 code implementation17 Mar 2022 Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, Hao Su

We demonstrate that applying traditional CP decomposition -- that factorizes tensors into rank-one components with compact vectors -- in our framework leads to improvements over vanilla NeRF.

PINA: Learning a Personalized Implicit Neural Avatar from a Single RGB-D Video Sequence

no code implementations3 Mar 2022 Zijian Dong, Chen Guo, Jie Song, Xu Chen, Andreas Geiger, Otmar Hilliges

We present a novel method to learn Personalized Implicit Neural Avatars (PINA) from a short RGB-D sequence.

Frame

StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasets

1 code implementation1 Feb 2022 Axel Sauer, Katja Schwarz, Andreas Geiger

StyleGAN in particular sets new standards for generative modeling regarding image quality and controllability.

Image Generation

gDNA: Towards Generative Detailed Neural Avatars

no code implementations11 Jan 2022 Xu Chen, Tianjian Jiang, Jie Song, Jinlong Yang, Michael J. Black, Andreas Geiger, Otmar Hilliges

Furthermore, we show that our method can be used on the task of fitting human models to raw scans, outperforming the previous state-of-the-art.

RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputs

no code implementations1 Dec 2021 Michael Niemeyer, Jonathan T. Barron, Ben Mildenhall, Mehdi S. M. Sajjadi, Andreas Geiger, Noha Radwan

We observe that the majority of artifacts in sparse input scenarios are caused by errors in the estimated scene geometry, and by divergent behavior at the start of training.

Novel View Synthesis

On the Frequency Bias of Generative Models

1 code implementation NeurIPS 2021 Katja Schwarz, Yiyi Liao, Andreas Geiger

2) Checkerboard artifacts introduced by upsampling cannot explain the spectral discrepancies alone as the generator is able to compensate for these artifacts.

Projected GANs Converge Faster

3 code implementations NeurIPS 2021 Axel Sauer, Kashyap Chitta, Jens Müller, Andreas Geiger

Generative Adversarial Networks (GANs) produce high-quality images but are challenging to train.

Image Generation

ATISS: Autoregressive Transformers for Indoor Scene Synthesis

no code implementations NeurIPS 2021 Despoina Paschalidou, Amlan Kar, Maria Shugrina, Karsten Kreis, Andreas Geiger, Sanja Fidler

The ability to synthesize realistic and diverse indoor furniture layouts automatically or based on partial input, unlocks many applications, from better interactive 3D tools to data synthesis for training and simulation.

Indoor Scene Synthesis

KITTI-360: A Novel Dataset and Benchmarks for Urban Scene Understanding in 2D and 3D

2 code implementations28 Sep 2021 Yiyi Liao, Jun Xie, Andreas Geiger

For the last few decades, several major subfields of artificial intelligence including computer vision, graphics, and robotics have progressed largely independently from each other.

Scene Understanding Self-Driving Cars

NEAT: Neural Attention Fields for End-to-End Autonomous Driving

1 code implementation ICCV 2021 Kashyap Chitta, Aditya Prakash, Andreas Geiger

Efficient reasoning about the semantic, spatial, and temporal structure of a scene is a crucial prerequisite for autonomous driving.

Autonomous Driving Imitation Learning

Learning Cascaded Detection Tasks with Weakly-Supervised Domain Adaptation

no code implementations9 Jul 2021 Niklas Hanselmann, Nick Schneider, Benedikt Ortelt, Andreas Geiger

In order to handle the challenges of autonomous driving, deep learning has proven to be crucial in tackling increasingly complex tasks, such as 3D detection or instance segmentation.

Autonomous Driving Domain Adaptation +2

MetaAvatar: Learning Animatable Clothed Human Models from Few Depth Images

1 code implementation NeurIPS 2021 Shaofei Wang, Marko Mihajlovic, Qianli Ma, Andreas Geiger, Siyu Tang

In contrast, we propose an approach that can quickly generate realistic clothed human avatars, represented as controllable neural SDFs, given only monocular depth images.

Meta-Learning

Shape As Points: A Differentiable Poisson Solver

1 code implementation NeurIPS 2021 Songyou Peng, Chiyu "Max" Jiang, Yiyi Liao, Michael Niemeyer, Marc Pollefeys, Andreas Geiger

However, the implicit nature of neural implicit representations results in slow inference time and requires careful initialization.

3D Reconstruction Surface Reconstruction

Locally Aware Piecewise Transformation Fields for 3D Human Mesh Registration

no code implementations CVPR 2021 Shaofei Wang, Andreas Geiger, Siyu Tang

We combine PTF with multi-class occupancy networks, obtaining a novel learning-based framework that learns to simultaneously predict shape and per-point correspondences between the posed space and the canonical space for clothed human.

Surface Reconstruction Translation

SMD-Nets: Stereo Mixture Density Networks

1 code implementation CVPR 2021 Fabio Tosi, Yiyi Liao, Carolin Schmitt, Andreas Geiger

Despite stereo matching accuracy has greatly improved by deep learning in the last few years, recovering sharp boundaries and high-resolution outputs efficiently remains challenging.

Disparity Estimation Stereo Matching

SNARF: Differentiable Forward Skinning for Animating Non-Rigid Neural Implicit Shapes

1 code implementation ICCV 2021 Xu Chen, Yufeng Zheng, Michael J. Black, Otmar Hilliges, Andreas Geiger

However, this is problematic since the backward warp field is pose dependent and thus requires large amounts of data to learn.

CAMPARI: Camera-Aware Decomposed Generative Neural Radiance Fields

no code implementations31 Mar 2021 Michael Niemeyer, Andreas Geiger

At test time, our model generates images with explicit control over the camera as well as the shape and appearance of the scene.

3D-Aware Image Synthesis

KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs

2 code implementations ICCV 2021 Christian Reiser, Songyou Peng, Yiyi Liao, Andreas Geiger

NeRF synthesizes novel views of a scene with unprecedented quality by fitting a neural radiance field to RGB images.

Neural Parts: Learning Expressive 3D Shape Abstractions with Invertible Neural Networks

1 code implementation CVPR 2021 Despoina Paschalidou, Angelos Katharopoulos, Andreas Geiger, Sanja Fidler

The INN allows us to compute the inverse mapping of the homeomorphism, which in turn, enables the efficient computation of both the implicit surface function of a primitive and its mesh, without any additional post-processing.

Counterfactual Generative Networks

1 code implementation ICLR 2021 Axel Sauer, Andreas Geiger

Prior works on image classification show that instead of learning a connection to object shape, deep classifiers tend to exploit spurious correlations with low-level texture or the background for solving the classification task.

Classification General Classification +2

GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields

1 code implementation CVPR 2021 Michael Niemeyer, Andreas Geiger

While several recent works investigate how to disentangle underlying factors of variation in the data, most of them operate in 2D and hence ignore that our world is three-dimensional.

Image Generation Neural Rendering

Category Level Object Pose Estimation via Neural Analysis-by-Synthesis

no code implementations ECCV 2020 Xu Chen, Zijian Dong, Jie Song, Andreas Geiger, Otmar Hilliges

Many object pose estimation algorithms rely on the analysis-by-synthesis framework which requires explicit representations of individual object instances.

Image Generation Pose Estimation

GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis

1 code implementation NeurIPS 2020 Katja Schwarz, Yiyi Liao, Michael Niemeyer, Andreas Geiger

In contrast to voxel-based representations, radiance fields are not confined to a coarse discretization of the 3D space, yet allow for disentangling camera and scene properties while degrading gracefully in the presence of reconstruction ambiguity.

3D-Aware Image Synthesis Novel View Synthesis +1

Intrinsic Autoencoders for Joint Neural Rendering and Intrinsic Image Decomposition

no code implementations29 Jun 2020 Hassan Abu Alhaija, Siva Karthik Mustikovela, Justus Thies, Varun Jampani, Matthias Nießner, Andreas Geiger, Carsten Rother

Neural rendering techniques promise efficient photo-realistic image synthesis while at the same time providing rich control over scene parameters by learning the physical image formation process.

Image-to-Image Translation Intrinsic Image Decomposition +1

Benchmarking Unsupervised Object Representations for Video Sequences

1 code implementation12 Jun 2020 Marissa A. Weis, Kashyap Chitta, Yash Sharma, Wieland Brendel, Matthias Bethge, Andreas Geiger, Alexander S. Ecker

Perceiving the world in terms of objects and tracking them through time is a crucial prerequisite for reasoning and scene understanding.

Multi-Object Tracking Object Detection +1

Learning Neural Light Transport

no code implementations5 Jun 2020 Paul Sanzenbacher, Lars Mescheder, Andreas Geiger

In recent years, deep generative models have gained significance due to their ability to synthesize natural-looking images with applications ranging from virtual reality to data augmentation for training computer vision models.

Data Augmentation Image Denoising

Label Efficient Visual Abstractions for Autonomous Driving

3 code implementations20 May 2020 Aseem Behl, Kashyap Chitta, Aditya Prakash, Eshed Ohn-Bar, Andreas Geiger

Beyond label efficiency, we find several additional training benefits when leveraging visual abstractions, such as a significant reduction in the variance of the learned policy when compared to state-of-the-art end-to-end driving models.

Autonomous Driving Semantic Segmentation

Learning Unsupervised Hierarchical Part Decomposition of 3D Objects from a Single RGB Image

1 code implementation CVPR 2020 Despoina Paschalidou, Luc van Gool, Andreas Geiger

Humans perceive the 3D world as a set of distinct objects that are characterized by various low-level (geometry, reflectance) and high-level (connectivity, adjacency, symmetry) properties.

3D Reconstruction

Learning Implicit Surface Light Fields

2 code implementations27 Mar 2020 Michael Oechsle, Michael Niemeyer, Lars Mescheder, Thilo Strauss, Andreas Geiger

In this work, we propose a novel implicit representation for capturing the visual appearance of an object in terms of its surface light field.

3D Reconstruction Image Generation

Convolutional Occupancy Networks

4 code implementations ECCV 2020 Songyou Peng, Michael Niemeyer, Lars Mescheder, Marc Pollefeys, Andreas Geiger

Recently, implicit neural representations have gained popularity for learning-based 3D reconstruction.

3D Reconstruction

Self-Supervised Linear Motion Deblurring

1 code implementation10 Feb 2020 Peidong Liu, Joel Janai, Marc Pollefeys, Torsten Sattler, Andreas Geiger

Motion blurry images challenge many computer vision algorithms, e. g, feature detection, motion estimation, or object recognition.

Deblurring Image Deblurring +3

Towards Unsupervised Learning of Generative Models for 3D Controllable Image Synthesis

1 code implementation CVPR 2020 Yiyi Liao, Katja Schwarz, Lars Mescheder, Andreas Geiger

We define the new task of 3D controllable image synthesis and propose an approach for solving it by reasoning both in 3D space and in the 2D image domain.

Image Generation

Attacking Optical Flow

1 code implementation ICCV 2019 Anurag Ranjan, Joel Janai, Andreas Geiger, Michael J. Black

In this paper, we extend adversarial patch attacks to optical flow networks and show that such attacks can compromise their performance.

Optical Flow Estimation Self-Driving Cars

Texture Fields: Learning Texture Representations in Function Space

no code implementations ICCV 2019 Michael Oechsle, Lars Mescheder, Michael Niemeyer, Thilo Strauss, Andreas Geiger

A major reason for these limitations is that common representations of texture are inefficient or hard to interface for modern deep learning techniques.

Robust Dense Mapping for Large-Scale Dynamic Environments

no code implementations7 May 2019 Ioan Andrei Bârsan, Peidong Liu, Marc Pollefeys, Andreas Geiger

We use both instance-aware semantic segmentation and sparse scene flow to classify objects as either background, moving, or potentially moving, thereby ensuring that the system is able to model objects with the potential to transition from static to dynamic, such as parked cars.

Semantic Segmentation Visual Odometry

Superquadrics Revisited: Learning 3D Shape Parsing beyond Cuboids

1 code implementation CVPR 2019 Despoina Paschalidou, Ali Osman Ulusoy, Andreas Geiger

Abstracting complex 3D shapes with parsimonious part-based representations has been a long standing goal in computer vision.

reinforcement-learning

RayNet: Learning Volumetric 3D Reconstruction with Ray Potentials

1 code implementation CVPR 2018 Despoina Paschalidou, Ali Osman Ulusoy, Carolin Schmitt, Luc van Gool, Andreas Geiger

RayNet integrates a CNN that learns view-invariant feature representations with an MRF that explicitly encodes the physics of perspective projection and occlusion.

3D Reconstruction

Taking a Deeper Look at the Inverse Compositional Algorithm

1 code implementation CVPR 2019 Zhaoyang Lv, Frank Dellaert, James M. Rehg, Andreas Geiger

In this paper, we provide a modern synthesis of the classic inverse compositional algorithm for dense image alignment.

Motion Estimation

Geometric Image Synthesis

no code implementations12 Sep 2018 Hassan Abu Alhaija, Siva Karthik Mustikovela, Andreas Geiger, Carsten Rother

The task of generating natural images from 3D scenes has been a long standing goal in computer graphics.

Image Generation Instance Segmentation +1

Learning Priors for Semantic 3D Reconstruction

no code implementations ECCV 2018 Ian Cherabier, Johannes L. Schonberger, Martin R. Oswald, Marc Pollefeys, Andreas Geiger

In contrast to existing variational methods for semantic 3D reconstruction, our model is end-to-end trainable and captures more complex dependencies between the semantic labels and the 3D geometry.

3D Reconstruction

SphereNet: Learning Spherical Representations for Detection and Classification in Omnidirectional Images

no code implementations ECCV 2018 Benjamin Coors, Alexandru Paul Condurache, Andreas Geiger

Omnidirectional cameras offer great benefits over classical cameras wherever a wide field of view is essential, such as in virtual reality applications or in autonomous robots.

General Classification Image Classification +1

Conditional Affordance Learning for Driving in Urban Environments

1 code implementation18 Jun 2018 Axel Sauer, Nikolay Savinov, Andreas Geiger

Most existing approaches to autonomous driving fall into one of two categories: modular pipelines, that build an extensive model of the environment, and imitation learning approaches, that map images directly to control outputs.

Autonomous Driving Autonomous Navigation +1

Learning 3D Shape Completion From Laser Scan Data With Weak Supervision

1 code implementation CVPR 2018 David Stutz, Andreas Geiger

Learning-based approaches, in contrast, avoid the expensive optimization step and instead directly predict the complete shape from the incomplete observations using deep neural networks.

Deep Marching Cubes: Learning Explicit Surface Representations

1 code implementation CVPR 2018 Yiyi Liao, Simon Donné, Andreas Geiger

Existing learning based solutions to 3D surface prediction cannot be trained end-to-end as they operate on intermediate representations (e. g., TSDF) from which 3D surface meshes must be extracted in a post-processing step (e. g., via the marching cubes algorithm).

Learning 3D Shape Completion under Weak Supervision

4 code implementations18 May 2018 David Stutz, Andreas Geiger

We address the problem of 3D shape completion from sparse and noisy point clouds, a fundamental problem in computer vision and robotics.

Which Training Methods for GANs do actually Converge?

9 code implementations ICML 2018 Lars Mescheder, Andreas Geiger, Sebastian Nowozin

In this paper, we show that the requirement of absolute continuity is necessary: we describe a simple yet prototypical counterexample showing that in the more realistic case of distributions that are not absolutely continuous, unregularized GAN training is not always convergent.

On the Integration of Optical Flow and Action Recognition

no code implementations22 Dec 2017 Laura Sevilla-Lara, Yiyi Liao, Fatma Guney, Varun Jampani, Andreas Geiger, Michael J. Black

Here we take a deeper look at the combination of flow and action recognition, and investigate why optical flow is helpful, what makes a flow method good for action recognition, and how we can make it better.

Action Recognition Optical Flow Estimation

Semantic Visual Localization

no code implementations CVPR 2018 Johannes L. Schönberger, Marc Pollefeys, Andreas Geiger, Torsten Sattler

Robust visual localization under a wide range of viewing conditions is a fundamental problem in computer vision.

Visual Localization

Sparsity Invariant CNNs

1 code implementation22 Aug 2017 Jonas Uhrig, Nick Schneider, Lukas Schneider, Uwe Franke, Thomas Brox, Andreas Geiger

In this paper, we consider convolutional neural networks operating on sparse inputs with an application to depth upsampling from sparse laser scan data.

Depth Completion Depth Estimation

Augmented Reality Meets Computer Vision : Efficient Data Generation for Urban Driving Scenes

no code implementations4 Aug 2017 Hassan Abu Alhaija, Siva Karthik Mustikovela, Lars Mescheder, Andreas Geiger, Carsten Rother

Further, we demonstrate the utility of our approach on training standard deep models for semantic instance segmentation and object detection of cars in outdoor driving scenes.

Instance Segmentation Object Detection +1

Semantic Multi-View Stereo: Jointly Estimating Objects and Voxels

no code implementations CVPR 2017 Ali Osman Ulusoy, Michael J. Black, Andreas Geiger

Due to its probabilistic nature, the approach is able to cope with the approximate geometry of the 3D models as well as input shapes that are not present in the scene.

3D Reconstruction

Toroidal Constraints for Two-Point Localization Under High Outlier Ratios

no code implementations CVPR 2017 Federico Camposeco, Torsten Sattler, Andrea Cohen, Andreas Geiger, Marc Pollefeys

Adding the knowledge of direction of triangulation, we are able to approximate the position of the camera from two matches alone.

Pose Estimation

The Numerics of GANs

4 code implementations NeurIPS 2017 Lars Mescheder, Sebastian Nowozin, Andreas Geiger

In this paper, we analyze the numerics of common algorithms for training Generative Adversarial Networks (GANs).

Computer Vision for Autonomous Vehicles: Problems, Datasets and State of the Art

no code implementations18 Apr 2017 Joel Janai, Fatma Güney, Aseem Behl, Andreas Geiger

Towards this goal, we analyze the performance of the state of the art on several challenging benchmarking datasets, including KITTI, MOT, and Cityscapes.

Autonomous Driving Motion Estimation +1

OctNetFusion: Learning Depth Fusion from Data

1 code implementation4 Apr 2017 Gernot Riegler, Ali Osman Ulusoy, Horst Bischof, Andreas Geiger

In this paper, we present a learning based approach to depth fusion, i. e., dense 3D reconstruction from multiple depth images.

3D Reconstruction

Adversarial Variational Bayes: Unifying Variational Autoencoders and Generative Adversarial Networks

1 code implementation ICML 2017 Lars Mescheder, Sebastian Nowozin, Andreas Geiger

We show that in the nonparametric limit our method yields an exact maximum-likelihood assignment for the parameters of the generative model, as well as the exact posterior distribution over the latent variables given an observation.

Probabilistic Duality for Parallel Gibbs Sampling without Graph Coloring

no code implementations21 Nov 2016 Lars Mescheder, Sebastian Nowozin, Andreas Geiger

We present a new notion of probabilistic duality for random variables involving mixture distributions.

Exploiting Object Similarity in 3D Reconstruction

no code implementations ICCV 2015 Chen Zhou, Fatma Guney, Yizhou Wang, Andreas Geiger

Despite recent progress, reconstructing outdoor scenes in 3D from movable platforms remains a highly difficult endeavour.

3D Reconstruction Frame

Displets: Resolving Stereo Ambiguities Using Object Knowledge

no code implementations CVPR 2015 Fatma Guney, Andreas Geiger

Stereo techniques have witnessed tremendous progress over the last decades, yet some aspects of the problem still remain challenging today.

Semantic Segmentation

Object Scene Flow for Autonomous Vehicles

no code implementations CVPR 2015 Moritz Menze, Andreas Geiger

We demonstrate the performance of our model on existing benchmarks as well as a novel realistic dataset with scene flow ground truth.

Autonomous Driving Scene Flow Estimation +1

Lost! Leveraging the Crowd for Probabilistic Visual Self-Localization

no code implementations CVPR 2013 Marcus A. Brubaker, Andreas Geiger, Raquel Urtasun

In this paper we propose an affordable solution to selflocalization, which utilizes visual odometry and road maps as the only inputs.

Visual Odometry

Joint 3D Estimation of Objects and Scene Layout

no code implementations NeurIPS 2011 Andreas Geiger, Christian Wojek, Raquel Urtasun

We propose a novel generative model that is able to reason jointly about the 3D scene layout as well as the 3D location and orientation of objects in the scene.

Cannot find the paper you are looking for? You can Submit a new open access paper.