25 lines
2.8 KiB
TeX
25 lines
2.8 KiB
TeX
With the advances in 3D models editing and 3D reconstruction techniques, more and more 3D models are available and their quality is increasing.
|
|
Furthermore, the support of 3D visualisation on the web has become standard during the last years.
|
|
A major problem that emerges from this situation is how users visualise and interact with these new high quality models and how to deliver them to remote users.
|
|
This thesis focuses on streaming and interaction of 3D remote virtual environments, and describes three major contributions.
|
|
|
|
First, we propose a 3D scene navigation interface with bookmarks, which are small virtual objects added to the scene that the user can click to easily move towards a recommended location.
|
|
We describe a user study where participants can navigate in 3D scenes with and without bookmarks.
|
|
Users navigate and accomplish the task faster if the virtual environment contains bookmarks.
|
|
However, this faster navigation has a drawback in the streaming scenario: a user who moves faster in a scene requires higher streaming capabilities in order to have the same quality of service.
|
|
This drawback can be mitigated due to the fact that bookmarks positions are known in advance: we propose a way of precomputing visible faces from a bookmark and use it during the streaming to decrease the latency when users click on bookmarks.
|
|
|
|
Secondly, we propose an adaptation of Dynamic Adaptive Streaming over HTTP (DASH), the video streaming standard, to 3D streaming.
|
|
To do so, we cut the scene into a k-d tree where each cell correspond to a DASH adaptation set.
|
|
Each cell is further divided into segments of a fixed number of faces, grouping together faces of similar areas.
|
|
Each texture is stored in its own adaptation set, and multiple representations are available for different resolutions of the textures.
|
|
All the metadata (the cells of the k-d tree, the resolutions of the textures, etc...) is encoded in an XML file.
|
|
We then propose a client capable of evaluating the usefulness of each chunk of data, and a few streaming policies that decide which chunks to download.
|
|
|
|
Finally, we investigate the case of 3D streaming and navigation on mobile devices.
|
|
We propose interfaces for 3D navigation on both desktop and mobile devices, with a streaming system based on DASH.
|
|
We integrate bookmarks in our 3D version of DASH and we propose a more adapted version of the precomputations described earlier.
|
|
We describe another user study where participants navigated in a 3D scene with a mobile device, with or without bookmarks, and with or without using the precomputation to enhance streaming when users click on bookmarks.
|
|
Out of 18 participants, 10 preferred our optimized policy, 4 preferred the default version and 4 had no opinion.
|
|
This seems to indicate that by augmenting a 3D interface with bookmarks, we are able to increase both the quality of service and the quality of experience of the users.
|