Distributed 3D model optimization for the web with the common implementation framework for online virtual museums

Internet services are becoming more ubiquitous and 3D graphics is increasingly gaining a strong foothold in the Web technology domain. Recently, with WebGL, real-time 3D graphics in the Browser became a reality and most major Browsers support WebGL natively today. This makes it possible to create applications like 3D catalogs of artifacts, or to interactively explore Cultural Heritage objects in a Virtual Museum on mobile devices. Frameworks like the open-source system X3DOM provide declarative access to low-level GPU routines along with seamless integration of 3D graphics into HTML5 applications through standardized Web technologies. Most 3D models also need to be optimized to address concerns like limited network bandwidth or reduced GPU power on mobile devices. Therefore, recently an online platform for the development of Virtual Museums with particular attention to presentation and visualization of Cultural Heritage assets in online virtual museums was proposed [1]. This common implementation Framework (CIF) allows the user to upload large 3D models, which are subsequently converted and optimized for web display and embedded in an HTML5 application that can range from simple interactive display of the model to an entire virtual environment like a virtual walk-through. Generating these various types of applications is done via a templating mechanism, which will be further elaborated within this paper. Moreover, to efficiently convert many large models into an optimized form, a substantial amount of computing power is required, which a single system cannot yet provide in a timely fashion. Therefore, we also describe how the CIF can be used to utilize a dynamically allocated cloud-based or physical cluster of commodity hardware to distribute the workload of model optimization for the Web.