Publications


2023


  1. Combining Resampled Importance and Projected Solid Angle Samplings for Many Area Light Rendering

    Ishaan Shah*Aakash KT*, and P. J. Narayanan
    ACM SIGGRAPH Asia 2023, Technical Communications (Formerly Technical Briefs)

    We identify the core issue for the high run-times of a naive combination of RIS and projected solid angle sampling. We then reformulate RIS for a better integration with projected solid angle sampling, achieving the state-of-the-art in many area light rendering.


  2. Analytical & Neural approaches to Physically Based Rendering

    Aakash KT
    ACM SIGGRAPH Asia 2023, Doctoral Consortium

    Summary of my PhD thesis, including two extensions to LTCs and the extension of NRC for efficient hair rendering.


  3. Accelerating Hair Rendering by Learning High-Order Scattered Radiance

    Eurographics Symposium on Rendering (EGSR) 2023, Computer Graphics Forum (CGF) Vol. 42, No. 4

    Efficiently and accurately rendering hair accounting for multiple scattering is a challenging open problem. We present a technique to infer the higher order scattering in hair in constant time within the path tracing framework, while achieving better computational efficiency.



2022


  1. Real-Time Rendering of Arbitrary Surface Geometries using Learnt Transfer

    Sirikonda DhawalAakash KT, and P. J. Narayanan
    ICVGIP 2022, Full Paper

    In this paper, we propose a compact transfer representation that is learnt directly on scene geometry points. Specifically, we train a small multi-layer perceptron (MLP) to predict the transfer at sampled surface points.


  2. Learnt Transfer for Surface Geometries

    Sirikonda DhawalAakash KT, and P. J. Narayanan
    High Performance Graphics 2022 (HPG), Poster


  3. Bringing Linearly Transformed Cosines to Anisotropic GGX

    ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games (I3D) 2022, PACM-CGIT, Best Paper Award

    We present an extension to LTCs for anisotropic GGX, in the context of real-time analytic area light shading. Our approach robustly fits LTC matrices to anisotropic GGX and ensures artefact free rendering. The end result is a 84 LUT parameterized by the elevation and azimuth of view vector and the roughness in x and y directions.


  4. PRTT: Precomputed Radiance Transfer Textures

    Sirikonda DhawalAakash KT, and P. J. Narayanan
    Eurographics (EG) 2022, Poster

    We analyze and extend PRT to use textures for storing transfer, instead of at vertices of a mesh. We demonstrate better rendering quality for the same mesh resolution for glossy reflection and inter-reflections. We also analyze the run-time, memory requirements and demonstrate benefits of using transfer textures.



2021


  1. Fast Analytic Soft Shadows from Area Lights

    Eurographics Symposium on Rendering (EGSR) 2021, DL-only track

    We present an analytical solution for soft shadows from area lights, which naturally produces noise-free renderings as compared to equivalent stochastic methods. A structured approach to analytically compute soft shadows from spherical projections of lights and occluders with any 3D shape and efficiently for convex 3D shapes.


  2. Neural View Synthesis with Appearance Editing from Unstructured Images

    ICVGIP 2021, Full Paper

    We present a neural rendering framework for simultaneous view synthesis and appearance editing of a scene with known environmental illumination captured using a mobile camera. Our approach explicitly disentangles the appearance and learns a lighting representation that is independent of it. We show results of editing the appearance of real scenes in interesting and non-trivial ways.



2019


  1. Exploring Data Driven Graphics for Interactivity

    Aakash KT, and P. J. Narayanan
    IIIT-H, Master’s Thesis


  2. A Flexible Neural Renderer for Material Visualization

    ACM SIGGRAPH Asia 2019, Technical Briefs

    Photo realism in computer generated imagery is crucially dependent on how well an artist is able to recreate real-world materials in the scene. We propose a convolutional neural network based workflow which generates high-quality ray traced material visualizations on a shaderball, in real-time.