**5. Conclusions and acknowledgements**

For using a photorealistic approach in 3D telepresence, the biggest challenges are the accuracy and cost of acquiring, delivering, and displaying 3D captures. Glasses based approaches are attractive due to their ability to support user mobility, immersion, and in future even natural focus/accommodation in 3D perception. Glasses can provide also sensors for 3D capture and user positioning. However, natural occlusions and accommodation are hard to support, and likely it will still take many years before affordable and good-enough glasses are available.

In this article, in addition to reviewing the general status and approaches for 3D telepresence, we proposed a simplified approach for spatially faithful telepresence, based on 3D screens and low bitrate streaming of video-plus-depth data. Screen displays are cheap, and can support 3D cues and mobility better than in existing teleconferencing solutions. External displays are easy to view and are a good option for improved 3D telepresence solutions. An important enabler for a simplified solution is efficient coding and streaming of RGB-D data, for which an exemplary implementation was presented with some simulation results. Described means for 3D capture and streaming support also XR functionalities, which in addition to viewing XR content on screens, may also be used with glasses or mobile displays. Features of the introduced solution can be summarized as follows:

	- 1. replacing a participant capture by an avatar, and
	- 2. replacing a captured meeting environment by a virtual space

Choosing screen displays implies both limitations and benefits to the system. With one or few screen displays, the perception of a meeting geometry is not same between participants. We decided to rely on the benefits of an exaggerated perception for eye contact by collecting all remote participant renderings into a grid on (nominally) one screen display, and providing viewers with perception for depth and motion parallax.

The key enabler for the improved system is simple: support for real-time capture, coding and streaming of video-plus-depth data from the RGB-D sensor of a telepresence terminal. This choice enables low bitrates, depth perception, and support for small viewpoint changes by user motions. Currently, the most feasible option for a display is a stereoscopic (S3D) display, but it can be replaced by an accommodation supportive display as soon they come available.

Our basic assumption is that each participant has his/her own telepresence terminal in the same way as PCs and laptops are currently used in videoconferencing. Correspondingly, from appearance, the new solution does not differ too much from current videoconferencing systems. For better supporting spatial faithfulness and gaze awareness, a more complicated system with several cameras and 3D screen displays may be used (i.e. applying Hydra and Viewport type of approaches introduced in Chapter 2.1). In parallel to viewing remote participants on screens, AR glasses can be used for viewing augmented objects inside a local space.

Although we have tested only some important prerequisites of the suggested solution, we have now a good knowledge and plan for its full implementation. We hope that this article raises its readers a general interest to the status, challenges, and possibilities of 3D telepresence, as well as a specific interest to develop the described approaches and ideas even further.

We want to express our gratitude to VTT Technical Research Centre of Finland for giving us the opportunity to work on 3D telepresence and related topics. Thanks also for InterDigital Inc. for challenging us to make new inventions in this area.

*Advances in Spatially Faithful (3D) Telepresence DOI: http://dx.doi.org/10.5772/intechopen.99271*
