Key papers in Computer Graphics and Computer-Aided Design

Pofolia’s corpus holds 79 papers from the Computer Graphics and Computer-Aided Design subfield (2012–2024). The list below starts with the most cited.

Most cited

Ranked by citation count. Because citations accumulate over time, this list naturally leans towards work published a few years ago; for where the field is now, see “recently added”.

  • The Design of Everyday Things

    Vahlen eBooks · 2016 · FWCI 59.17 · 6,416 citations

    Product design, not user error, causes frustration with everyday objects; good design leverages visibility, natural mappings, and constraints to guide users effortlessly.

    Go to source

  • NeRF

    Communications of the ACM · 2021 · Q1 · SJR 1.00 · FWCI 975.12 · 5,949 citations · Open access

    A new method uses a deep neural network to represent complex 3D scenes as a continuous volumetric function, enabling photorealistic novel view synthesis from sparse input images.

    Go to source

  • Instant neural graphics primitives with a multiresolution hash encoding

    ACM Transactions on Graphics · 2022 · Q1 · SJR 5.00 · FWCI 1104.42 · 3,739 citations · Open access

    A novel multiresolution hash encoding significantly speeds up neural graphics primitive training and rendering by using a small neural network augmented with a trainable hash table.

    Go to source

  • Screened poisson surface reconstruction

    ACM Transactions on Graphics · 2013 · Q1 · SJR 5.00 · FWCI 143.12 · 2,226 citations

    This work extends Poisson surface reconstruction by incorporating points as explicit interpolation constraints, generalizing the mathematical framework to a screened Poisson equation.

    Go to source

  • TetGen, a Delaunay-Based Quality Tetrahedral Mesh Generator

    ACM Transactions on Mathematical Software · 2015 · Q1 · FWCI 369.02 · 1,593 citations

    TetGen is a new C++ program that generates high-quality tetrahedral meshes for numerical methods and scientific computing using Delaunay-based algorithms.

    Go to source

  • Spatial Point Patterns

    2015 · FWCI 43.14 · 1,581 citations

    This book introduces modern statistical methods and R software for analyzing spatial point pattern data, making complex techniques accessible to researchers and statisticians across various fields.

    Go to source

  • D-NeRF: neural radiance fields for dynamic scenes

    UPCommons institutional repository (Universitat Politècnica de Catalunya) · 2021 · 1,252 citations

    This paper introduces D-NeRF, a novel method for reconstructing dynamic 3D scenes from a collection of images.

    Go to source

  • Efficient Geometry-aware 3D Generative Adversarial Networks

    2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · 2022 · 1,012 citations

    Researchers developed a new 3D GAN architecture that generates high-resolution, multi-view-consistent images and 3D shapes in real-time, overcoming limitations of previous methods.

    Go to source

  • 2021 IEEE/CVF International Conference on Computer Vision (ICCV)

    2021 · 986 citations

    SurfaceNet, a novel patch-based generative adversarial network (GAN), can estimate spatially-varying material properties (SVBRDF) from a single image, outperforming existing methods.

    Go to source

  • Structure extraction from texture via relative total variation

    ACM Transactions on Graphics · 2012 · Q1 · SJR 5.00 · FWCI 58.81 · 886 citations

    A new method using 'relative total variation' measures effectively extracts meaningful structures from textured surfaces, overcoming challenges posed by regular, near-regular, or irregular texture patterns.

    Go to source

  • READING IMAGES - THE GRAMMAR OF VISUAL DESIGN

    VNU Journal of Foreign Studies · 2017 · FWCI 176.89 · 861 citations · Open access

    This book introduces a systematic framework for analyzing visual images, treating them as a form of language with its own grammar.

    Go to source

  • Level of Detail for 3D Graphics

    2012 · FWCI 100.78 · 829 citations

    This book consolidates mechanisms, principles, and theory for Level of Detail (LOD) techniques, offering a practical resource for real-time graphics developers.

    Go to source

  • PlenOctrees for Real-time Rendering of Neural Radiance Fields

    2021 IEEE/CVF International Conference on Computer Vision (ICCV) · 2021 · 806 citations

    A new method, PlenOctrees, enables real-time rendering of Neural Radiance Fields (NeRFs) by pre-tabulating them into an octree structure, achieving over 150 FPS for 800x800 images.

    Go to source

  • Emerging MPEG Standards for Point Cloud Compression

    IEEE Journal on Emerging and Selected Topics in Circuits and Systems · 2018 · Q1 · FWCI 147.78 · 762 citations · Open access

    The Moving Picture Experts Group (MPEG) is developing a new open standard for compactly representing 3D point clouds, addressing the massive data requirements of immersive AR/VR experiences.

    Go to source

  • Magic3D: High-Resolution Text-to-3D Content Creation

    2023 · 732 citations

    Magic3D generates high-quality 3D mesh models from text prompts in 40 minutes, significantly faster and with higher resolution than previous methods like DreamFusion.

    Go to source

  • Real-Time Rendering

    2018 · FWCI 20.40 · 701 citations

    This fourth edition of 'Real-Time Rendering' is thoroughly updated to cover modern techniques for generating synthetic 3D images instantly, focusing on practical methods used in games and applications.

    Go to source

  • Shape, Illumination, and Reflectance from Shading

    IEEE Transactions on Pattern Analysis and Machine Intelligence · 2014 · Q1 · SJR 4.00 · FWCI 90.28 · 661 citations

    This paper introduces a novel statistical inference approach to recover 3D scene properties like shape, reflectance, and illumination from a single 2D image, overcoming the limitations of traditional multi-observation methods.

    Go to source

  • Learning hatching for pen-and-ink illustration of surfaces

    ACM Transactions on Graphics · 2012 · Q1 · SJR 5.00 · FWCI 88.21 · 649 citations

    A new algorithm learns to automatically generate pen-and-ink hatching illustrations in a specific artist's style by analyzing a single example drawing.

    Go to source

  • A dynamic seamless modeling method for the global multi-scale terrain based on DQG

    2015 · 601 citations

    A new method based on the spherical DQG (Degenerate Quadtree Grid) achieves dynamic, seamless global multi-scale terrain modeling, eliminating grid cracks and T-connections.

    Go to source

  • Radiosity and Realistic Image Synthesis

    2016 · FWCI 46.90 · 599 citations

    This paper introduces a framework for modeling radiative energy distribution in environments, such as tree crowns, to improve realistic image synthesis. The framework's formal specification allows for algorithm analysis and potential development of a specialized mini-language for environment modeling.

    Go to source

Recently added

  • 2D Gaussian Splatting for Geometrically Accurate Radiance Fields

    2024 · 474 citations

    2D Gaussian Splatting (2DGS) reconstructs geometrically accurate radiance fields by collapsing 3D Gaussians into view-consistent 2D planar Gaussian disks, overcoming 3DGS's surface representation limitations.

    Go to source

  • 4D Gaussian Splatting for Real-Time Dynamic Scene Rendering

    2024 · 569 citations

    4D Gaussian Splatting (4D-GS) achieves real-time dynamic scene rendering at 82 FPS (800x800 resolution on RTX 3090) by representing scenes with 3D Gaussians and 4D neural voxels, outperforming prior methods.

    Go to source

  • Mip-Splatting: Alias-Free 3D Gaussian Splatting

    2024 · 403 citations

    Mip-Splatting introduces a 3D smoothing filter and a Mip filter to eliminate aliasing artifacts in 3D Gaussian Splatting, enabling high-fidelity novel view synthesis across varying zoom levels.

    Go to source

  • Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering

    2024 · 328 citations

    Scaffold-GS introduces a novel approach to 3D Gaussian Splatting by using anchor points to dynamically predict Gaussian attributes based on viewing direction and distance, significantly reducing redundant Gaussians and improving robustness to view changes.

    Go to source

  • Wonder3D: Single Image to 3D Using Cross-Domain Diffusion

    2024 · 293 citations

    Wonder3D generates high-fidelity textured 3D meshes from single images in under 3 minutes by using a cross-domain diffusion model to create multi-view normal maps and color images, overcoming limitations of prior methods.

    Go to source

Add this field to your daily feed

Pick your interests and new work in your area arrives every day, summarised. Full summaries live in the app.

Open the app

Other subfields in the same field

All fields