Trivisio developed three different wearable displays. Two of them are stereoscopic optical see-through head-mounted displays (HMD). One HMD is using very bright state-of-the-art OLED microdisplays and offers a resolution of 320x240 pixels, the second HMD is using LCoS microdisplays with a resolution of 800x600 pixel. Both are very lightweight and showing very good image quality and offering a good wearing comfort. The final prototype is a hand held stereoscopic optical see-though binoculars offering pixel based real light blocking mechanism.
This innovative and unique approach offers groundbreaking possibilities for the planned application of architectural visualisation
Today, image transmission techniques are typically used for either streaming video or video conferencing systems.
In video streaming the video is usually compressed offline and transmitted on demand. In video conferencing the video images of the remote user are grabbed, compressed, transmitted, de-compressed and visualized on the receiver-side on-the-fly. Usually the resolution is low, the quality poor and there is a noticable lag.
Both approaches are not prepared for transmitting synchronized 3D stereoscopic images in near real-time.
The aim of this work-package is to develop means (based on standards such as MPEG) to provide real-time, high quality stereo visualization to a mobile end-user by transmitting over wireless networks realistically rendered images to his optical stereoscopic see-through HMD.
By rendering the content remotely on powerful graphics workstations and taking into account the current positional information of the mobile user, the traditional limitations of mobile computing power can be surpassed
The main goal of the mixed reality technology is to add computer-generated information to a real video sequence in such a manner that the real and virtual objects appear coexisting in the same world. In order to get a good illusion, the registration problem must be addressed. The real and virtual objects must be properly aligned with respect to each other. In this way, the position-orientation (pose) of the camera respect to a reference frame must be accurately estimated or updated over time. In IMPROVE, we address the registration problem for interactive mixed reality applications, based on a vision-based (optical) tracker working on a fully mobile wearable system.
ViconTech’s role in IMPROVE was to develop a low cost, markerless large-area tracking system that allows to determine the users head(camera) pose with 6 degrees of freedom. This transformation is estimated by processing the images acquired by the digital camera mounted on the Head Mounted Display. Our research focus on developing a markerless tracking component that allows to identify planar structures in unprepared environments, as robust and efficient as possible.
The purpose of a tracking system is to compute the position of a series of objects with respect to a fixed reference frame. These objects may be simply independent objetcs or may be related between them as in the case of the different segments of a human being. Optical tracking systems determine the 3D coordinates of some given points located on the objects and marked by reflective markers. The light reflected by the markers is seen by a set of cameras, which record the position of the markers in the camera reference. Based on this information and the data of the camera calibration the coordinates of the markers are calculated and from here the motion of the particular object is known. In order to compute the 3D coordinates of all the markers each marker must be seen at least by two cameras simultaneously. This is the origin of one of the main difficulties of the optical systems, which is the limited capture area that they can cover due to occlusions.
IMMIView is an interaction system that supports multi-modal and multi-user interaction. It was developed to offer a more natural interaction on Large Screen Displays. As input devices, it supports laser pointers, speech commands, body tracking and mobile device interaction. On Improve a novel stroke based interaction metaphors for Large Screen Displays was developed. This interface overcomes some laser interaction limitations and to provide a more natural interaction. Furthermore, other input devices, like speech recognition or body tracking, are also available to IMMIView, enabling multi-modal interaction. To support multi-modal interaction Inesc-ID developed an inference mechanism and a knowledge base that captures input and system information to better reason what multi-modal commands are being issues by the user. With this mechanism the prototype is able to issue multi-modal commands even with two or more active users.
IVIEW has been thought as a collaborative interface for design session within the automotive scenario. The user can define his/her own interaction dialogue by configuring IView through an external authoring tool. This includes the definition of specificic gestures, voice of standard menu-based command sequences to perform all functionalities supported by the application. This way the user is free to interact with the system in a completely multimodal manner by mixing and matching, gestures, calligraphical interfaces, standard menus or voice. The ideal layout sees two or more design reviewers sitting in front of a large tiled display system, each interacting through their personal IView interface. This can be done via voice, gestures (detected through a motion tracking system), ring menus, or calligraphical commands.
