Home Research Papers Datasets State of the Art Pricing

Discover, visualize, and connect AI research papers. Explore the latest trends and insights in artificial intelligence research.

Product

Home
Research Papers
About

Support

Contact
Terms of Service
Privacy Policy

© 2026 Papersgraph. All rights reserved.

computer-vision-2

Novel View Synthesis

3260 papers • 126 benchmarks • 313 datasets

Synthesize a target image with an arbitrary target camera pose from given source images and their camera poses. See Wiki for more introdcutions. The Synthesis method include: NeRF, MPI and so on. ( Image credit: Multi-view to Novel view: Synthesizing novel views with Self-Learned Confidence )

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in novel-view-synthesis-2

Trend

Dataset

Best Model

Actions

LLFF

LLFF

NeRF

NeRF

DONeRF: Evaluation Dataset

DONeRF: Evaluation Dataset

Libraries

i

Use these libraries to find novel-view-synthesis-2 models and implementations

ashawkey/torch-ngp

3 papers 1,995

Datasets

KITTI

NeRF

ShapeNet

SYNTHIA

LLFF

Mip-NeRF 360

KITTI-360

Subtasks

Novel LiDAR View Synthesis Gournd video synthesis from satellite image

Most implemented papers

Instant neural graphics primitives with a multiresolution hash encoding

Alex Evans, A. Keller, T. Müller, Christoph Schied•Sat Jan 15 2022

A versatile new input encoding that permits the use of a smaller network without sacrificing quality, thus significantly reducing the number of floating point and memory access operations is introduced, enabling training of high-quality neural graphics primitives in a matter of seconds, and rendering in tens of milliseconds at a resolution of 1920×1080.

5223

Content

Introduction Benchmarks Datasets Subtasks Libraries Papers

X3D

X3D

RTMV

RTMV

Tanks and Temples

Tanks and Temples

ACID

ACID

SWORD

SWORD

Mip-NeRF 360

Mip-NeRF 360

RealEstate10K

RealEstate10K

KITTI Novel View Synthesis

KITTI Novel View Synthesis

ShapeNet Car

ShapeNet Car

ShapeNet Chair

ShapeNet Chair

Synthia Novel View Synthesis

Synthia Novel View Synthesis

PhotoShape

PhotoShape

Dosovitskiy Chairs

Dosovitskiy Chairs

KITTI

KITTI

faceonlive/ai-research

3 papers 124

kakaobrain/NeRF-Factory

2 papers 1,243

ventusff/neurecon

2 papers 842

KITTI-360

BlendedMVS

Tanks and Temples

Tanks and Temples

Spaces

0

NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction

C. Theobalt, Lingjie Liu, Wenping Wang, T. Komura, Peng Wang, Yuan Liu•Sat Jun 19 2021

Experiments show that NeuS outperforms the state-of-the-arts in high-quality surface reconstruction, especially for objects and scenes with complex structures and self-occlusion.

2166 0

PlenOctrees for Real-time Rendering of Neural Radiance Fields

Matthew Tancik, Ruilong Li, Angjoo Kanazawa, Ren Ng, Hao Li, Alex Yu•Wed Mar 24 2021

It is shown that it is possible to train NeRFs to predict a spherical harmonic representation of radiance, removing the viewing direction as an input to the neural network, and PlenOctrees can be directly optimized to further minimize the reconstruction loss, which leads to equal or better quality compared to competing methods.

1244 0

View Synthesis by Appearance Flow

Alexei A. Efros, Jitendra Malik, Tinghui Zhou, Shubham Tulsiani, Weilun Sun•Tue May 10 2016

This work addresses the problem of novel view synthesis: given an input image, synthesizing new images of the same object or scene observed from arbitrary viewpoints and shows that for both objects and scenes, this approach is able to synthesize novel views of higher perceptual quality than previous CNN-based techniques.

721 0

HoloGAN: Unsupervised Learning of 3D Representations From Natural Images

Thu Nguyen-Phuoc, Chuan Li, Lucas Theis, Christian Richardt, Yong-Liang Yang•Mon Apr 01 2019

We propose a novel generative adversarial network (GAN) for the task of unsupervised learning of 3D representations from natural images. Most generative models rely on 2D kernels to generate images and make few assumptions about the 3D world. These models therefore tend to create blurry images or artefacts in tasks that require a strong 3D understanding, such as novel-view synthesis. HoloGAN instead learns a 3D representation of the world, and to render this representation in a realistic manner. Unlike other GANs, HoloGAN provides explicit control over the pose of generated objects through rigid-body transformations of the learnt 3D features. Our experiments show that using explicit 3D features enables HoloGAN to disentangle 3D pose and identity, which is further decomposed into shape and appearance, while still being able to generate images with similar or higher visual quality than other generative models. HoloGAN can be trained end-to-end from unlabelled 2D images only. Particularly, we do not require pose labels, 3D shapes, or multiple views of the same objects. This shows that HoloGAN is the first generative model that learns 3D representations from natural images in an entirely unsupervised manner.

541 0

SynSin: End-to-End View Synthesis From a Single Image

Justin Johnson, Georgia Gkioxari, R. Szeliski, Olivia Wiles•Tue Dec 17 2019

This work proposes a novel differentiable point cloud renderer that is used to transform a latent 3D point cloud of features into the target view and outperforms baselines and prior work on the Matterport, Replica, and RealEstate10K datasets.

520 0

Neural Body: Implicit Neural Representations with Structured Latent Codes for Novel View Synthesis of Dynamic Humans

H. Bao, Qing Shuai, Yuanqing Zhang, Xiaowei Zhou, Sida Peng, Yinghao Xu, Qianqian Wang•Wed Dec 30 2020

Neural Body is proposed, a new human body representation which assumes that the learned neural representations at different frames share the same set of latent codes anchored to a deformable mesh, so that the observations across frames can be naturally integrated.

817 0

Adding a benchmark result helps the community track progress.

Novel View Synthesis | State-of-the-Art