WebGPU Data Visualization Cookbook: (2nd Edition)

WebGPU is a state-of-the-art web-based hardware-acceleration interface for high performance graphics and compute. A emerging use of WebGPU is the development of real-time, high performance data visualization applications in fields ranging from medical imaging, simulations, engineering to cutting-edge extended reality computing. Indeed, data visualization has become increasingly challenging using conventional approaches without graphics hardware acceleration as datasets become larger and more complex, especially with the evolution of big data. WebGPU offers a flexible cross-platform solution for creating stunning visuals in 2D/3D in real-time for a wide range of interactive applications. This book contains a series of hands-on recipes that are tailored to both beginners who have very little experience with the WebGPU API and more advanced users who would like to explore state-of-the-art techniques. We begin with a basic introduction to the WebGPU API and WGSL shader language by demonstrating how to setup a basic project with outputs on screen. You'll then learn how to render basic 2D datasets with primitives, as well as more complex 3D volumetric datasets interactively. This part will focus on low-dimensional datasets that provide a spring-board for larger projects. After this, we transition to more advanced techniques, such as texture mapping for image/video processing, point cloud rendering of depth sensor data from 3D range-sensing cameras and stereoscopic 3D rendering. Also insights into eXtended Reality (XR); as we shift towards developing highly interactive XR applications using the WebXR API. Each recipe in this book gives readers a set of standard functions that can be imported to an existing project and can form the basis for the creation of a diverse array of real-time, interactive data visualization solutions. This book also utilizes a set of popular open-source libraries, such as `gl-matrix', to simplify application development and content management (loading data and models) as well as image/video processing using state-of-the-art compute vision algorithms.