Rendering Optimization for Mobile Web 3D Based on Animation Data Separation and On-Demand Loading

Abstract

Based on advances in image processing technology and Web-enabling technologies for mobile devices, mobile Augmented Reality (AR) and Virtual Reality (VR) has developed rapidly. The rendering and interaction of 3D models is an important part of AR and VR applications and is closely related to user experience. However, since the existing WebGL 3D JavaScript libraries for Web-based mobile 3D (represented by three.js and babylon.js) load the entire model file at once, large-size 3D models with complex interactions cannot be rendered smoothly due to limited data transmission, the weak computation capabilities of mobile Web browsers, and the latency of 3D model rendering. In this paper, we first propose model-animation data separation and an on-demand loading mechanism to improve the data request and loading process of Web 3D models. The main mechanisms are the following: (1) The model data are segmented into topological data and animation data sequences, and only the necessary data of the model are loaded when the Web-based mobile 3D model is first rendered. (2) The 3D model animation data sequence is semantically decomposed, and a multigranular model animation data service is established to provide continuous animation data support. (3) An asynchronous request-response mechanism is used to optimize the loading method of the model data. The model rendering mechanism uses an on-demand request and rendering method to transform the centralized loading process of the 3D model into a decentralized process. According to the testing and verification results, this optimization method can reduce the latency of mobile Web 3D in model data transmission and rendering by 24.72% for the experiment models. The interaction experience of Web-based mobile AR and VR is substantially improved relative to existing Web 3D rendering engines and rendering mechanisms, especially in complex interactive service scenarios. INDEX TERMS Mobile Web 3D, rendering interactive computing, on-demand loading, interfacing data services, augmented reality, virtual reality.

3B Görselleştirme Ortamı Olarak Web Tarayıcıları

“…WebGL frameworks such as Three.js or Babylon.js can be used with HTML5 in a development environment or JavaScript editors. They can load the model files at once [29]. Three.js is a Javascript library that allows 3D design on web browsers by benefit from the content of WebGL developed continuously.…”

A2W: Context-Aware Recommendation System for Mobile Augmented Reality Web Browser

“…AR applications can be categorised into two types: (1) Leveraging the web technologies for the implementation of platform-specific AR browsers [34,43], and (2) solely web browsers [20,36]. Although the existing works facilitate the AR content generation and rendering of static animated contents [32], the optimized design of AR layout, including visual content placement and user experience is neglected [24]. A2W parses a fully-functional webpage and achieves context-aware content placement in mobile headsets, and brings web contents in AR with user-centric perspective.…”

Enhanced LSTM-based Service Decomposition for Mobile Augmented Reality

“…However, note that the number of the total time steps equals the dimensions of the input matrix XTrain. It is followed by a dropout layer to avoid over-fitting by set a random number of elements to 0 (rand(size(X)) < P , X is a layer input and P is a self-defined probability) and scaling other elements by 1 1−P [8] [9]. In this design, two LSTM layers with their dropout layers are created to enhance the overall performance.…”

“…D IFFERENT from traditional applications, Mobile augmented reality (MAR) enhance virtual experiences through elevating capabilities of mobile devices and hence brings about significant flexibility and accuracy in amalgamating Augmented Reality(AR) objects with the physical world. However, it has naturally a higher degree of requirements in computing and caching resources, especially when rendering 3-dimensional (3D) AR objects [1]. To this end, Edge clouds (ECs) assistance allows the deployment of service at the edge of the network and hence is widely embedded in the design for orchestrating and offloading MAR services.…”

“…As illustrated in Fig. (1), the captured video of a MAR application is preprocessed at the terminal and then frames with AR objects are transmitted for detection, extraction and recognition. The local cache is searched to find if there is a match.…”