Real-time Three-Dimensional Diminished Reality (3DDR) develops a novel software framework for real-time diminished reality (DR), a variant of mixed reality (MR) which changes a user's real time perception so that unwanted objects disappear. Our new approach can do this in full 3D and in real time, so that users can move arbitrarily and see through or behind objects. Previous techniques have been limited to the assumption that the region in the image that must be diminished (replaced) is flat, which is often not the case in real-world applications. Our method restores a three-dimensional, non-flat background, and can also fill in the background if no previous observation exists. This type of filling -- a three-dimensional variant of so-called "inpainting" -- must be temporally coherent, so that perspective changes resulting from a moving user position do not destroy the illusion of DR. Our method lends itself to visual inspection tasks, aestethic improvements of mixed reality scenes and telepresence applications with high visual quality.
Fördergeber
- Fonds zur Förderung der wissenschaftlichen Forschung (FWF), FWF
Beginn: 31.12.2020
Instant Avatar - Instant Photorealistic Generation of Human Avatars with Mobile Devices
Capturing humans in every detail and displaying them in a photo-realistic way is one of the hardest challenges of computer graphics and computer vision. The human body shows a huge
variety in shape and appearance, and capturing every aspect of it is necessary to create a believable virtual representation of someone. Many technologies struggle with the so called
uncanny valley effect: This term describes the fact that “humanoid objects which appear almost, but not fully, like real human beings elicit revulsion in observers”.
State-of-the-art approaches in this area can generate digital human avatars, but do not reach photo-realism and fully believable results. Furthermore, current solutions require complex setups and powerful hardware to create a digital representation of a person. This limits the widespread use of the technology and therefore also impedes commercial success. Being able to overcome these problems would open up a large range of use cases. Applications in the fields of health care, education, e-commerce and entertainment, to only name a few, would benefit from such a technology. The activities of companies in this field emphasize the relevance of this topic. Almost every aspect of our digital communication will benefit from realistic avatars. To achieve this, the technical solution must be mobile and easy to use without the need for special hardware.
Thus, the project partners will combine 4D avatar capturing and image-based rendering in a system incorporating principles of photogrammetry and surface light fields to capture fine
details and view-dependent effects which are true to the original images of the subject. Facial expressions and body motion will be captured using novel tracking and classification algorithms
and can then be applied to avatars to create a realistic and believably human impression.
The Institute of Computer Graphics and Vision, lead of Prof. Dieter Schmalstieg, has decades of experience on the topics of 3D reconstruction, deep learning, augmented and virtual reality,
visualisation and other topics relevant to this project. A team of experts focused on research and development of mobile AR solutions and has released commercial products in this field.
Together with the funding of this project, all requirements to achieve both academic and commercial success are given.
Fördergeber
- Österreichische Forschungsförderungsgesellschaft mbH (FFG) , FFG
- Reactive Reality GmbH
Beginn: 30.11.2019
NNVRIK - Neural Network based Inverse Kinematics for Human Upper Body in Virtual Reality
Current consumer-grade virtual reality devices provide positional tracking for head and hands of the user. Using this information, virtual hands can be displayed to the user and by that be used for interaction in the virtual environment. However, other body parts are usually not visible to the user since there is no tracking information available. Furthermore, tracking information might be noisy or temporarily unavailable. In this project, we aim to close the gap between such incomplete tracking data and full body tracking, which might only become available in future consumer-grade virtual reality. Our aim is a combination of data driven (neural network-based) and inverse kinematics solvers for estimating joint data for the (upper body) pose of the user. Our previous work shows that, even if the pose is not correct, arms animated by inverse kinematics are preferred over having hands only and that they can also be used for interaction. Using neural networks and sufficient motion capture data, we expect to increase the accuracy even further and thus also increase the feeling of embodiment.
Fördergeber
- Facebook Technologies, LLC
Beginn: 30.04.2019
CAMed - Clinical Additive Manufacturing for Medical Applications
Additive manufacturing (AM) has developed rapidly into a technology enabling the creation of products with virtually limitless design geometries from an increasing number of materials spanning polymers, ceramics and metals. The innovative potential of AM is mirrored, in particular, by its impact on the medical manufacturing industry, which is supplying clinics with a steadily growing number of devices created by AM. The truly revolutionary and as yet completely unrealized potential of AM for medical applications lies, however, in the printing of
personalized implants directly in the clinic. Current treatment of fractures and/or lesions caused by trauma or tumour infiltration is often based on improvised usage of sub-optimal implants.
Complex pelvic fractures are, for example, fixed by intraoperative bending of standard metal plates to fit the fracture. The downsides of this are lengthy surgery and the creation of predetermined breaking points in the metal leading to later implant failure (and, thus, more surgery and attendant potential morbidity).
Ribs removed in the course of tumour resection are similarly replaced by relatively thin metal bands requiring intraoperative adaptation to patient specific curvature of the thorax. These
implants are, moreover, rigid in character and consequently eventually deformed by movement associated with breathing and the forces acting upon the thorax, necessitating
remedial surgery (often, multiple times).
Reconstruction following tumor-associated craniectomy automatically requires a second round of surgery because the customized implants used for this are currently produced by
external manufacturers (with a typical waiting time of six weeks). The potential benefits of customization in this case are, however, often forfeited due to changes to the shape of the
cranial opening that occur during the wait for the implant.
Additive manufacturing using optimized application-specific materials in the clinic itself would greatly improve overall outcome for these patients. It would, moreover, significantly reduce overall patient stress by obviating the need for a second surgical intervention (and others necessitated by possible complications).
Successful clinical implementation of AM will, however, necessitate the improvement and adaptation of key associated processes such as imaging, segmentation, material
development and the selection, in each case, of the most appropriate AM technology.
Subsequent developments will furthermore require effective integration of both the medical standpoint and the expertise of the (industrial) developers.
The CAMed project shall bring together clinicians, medical and material scientists and industry partners in a tightly networked, focused cooperation to develop AM-based processes for the clinical manufacture of custom-fit implants for specified medical
applications. Medical professionals and their patients will, as a result, begin to reap the enormous potential benefits of clinical 3D printing.
Fördergeber
- Medizinische Universität Graz
- Steirische Wirtschaftsförderungsgesellschaft m.b.H., SFG
- Österreichische Forschungsförderungsgesellschaft mbH (FFG) , FFG
Externe Partner
- JOANNEUM RESEARCH Forschungsgesellschaft mbH, MATERIALS - Institut für Oberflächentechnologien und Photonik
- Max-Planck-Institut für Intelligente Systeme
- Medizinische Universität Graz
- Montanuniversität Leoben, Institut für Werkstoffkunde und -prüfung der Kunststoffe
- Montanuniversität Leoben, Lehrstuhl für Kunststoffverarbeitung
Beginn: 31.10.2018
enFaced - Virtual and Augmented Reality for 3D Reconstruction
This interdisciplinary research project between computer science and medicine targets the development of a comprehensive image-guided tool for head and neck surgery with the main focus on mandibular fractures and mid-facial fractures. The tool supports the physicians in all treatment stages (diagnosis, planning, intervention and monitoring) and can also be used during clinical traineeship. A key point is the development of an algorithm for the semi-automatic and automatic segmentation of bone structures and soft tissue in CT and MRI acquisitions. A segmentation enables a three-dimensional localization, quantification and visualization of biological structures in a very short time. Based on the radiological images from the clinical routine, treatment diagnosis can be visualized, and surgical planning processes or alternative surgery options can be simulated. In addition, the postoperative simulation results can be presented preoperatively in a photo-realistic manner. In particular, this will be applied in the reposition of bony structures, the reconstruction of facial defects and the removal of tumors in complex anatomical areas. Moreover, we plan the development of patient-individual three-dimensional implants, which will 3D printed in-house for comparison with the external manufactured ones and for clinical usage. Furthermore, the medical personnel can be supported in real-time during surgery by an interactive navigation, where Augmented Reality integrates computer-generated objects into the operation field. The application reduces surgical complications and ensures a successful treatment result by pre- and intraoperative simulations. Hence, the number of necessary corrective surgeries and the operation time needed can be reduced, and a higher patient survival rate can be achieved. In addition, the combination with Virtual Reality glasses allows the virtual training of operations during medical education. The objective of the project is the establishment of an open source tool as basis for further research and developments for the participating universities. Currently, comparable tools are protected by strict licensing conditions and not accessible without restrictions for academic research. Existing software for clinical practice is not functionally stable enough and has too high error rates. Additionally, the usage imposes high financial costs, which makes a widespread application not possible. Finally, Augmented Reality in the clinical routine is an important topic for current surgical research and cannot be ignored by modern medical universities.
Fördergeber
- Fonds zur Förderung der wissenschaftlichen Forschung (FWF), FWF
Externe Partner
- Medizinische Universität Graz, Institut für Zellbiologie, Histologie und Embryologie
Beginn: 30.06.2018
Virtual Welding NEXT - Integration of Simulated and Real Welding
Through the research and development of innovative and efficient methods, a novel training simulator
shall be developed for welding, which enables a smooth transition between the virtual training and the
real environment, to improve the flexibility and usability of welding as core technology
in production in Industry 4.0.
The system prototype, including housing, burner, the helmet and the table) including the virtual welding processing
unit is to be remodelled as realistically as possible. The concept includes the design and development
of an intuitive graphical user interface to enable users the operation of the welding machine in a fast
and economic manner, at the same time enriching the know-how of the user and the understanding of safety and
working concepts.
The project Virtual Welding NEXT shall lower emotional barriers to use virtual and augmented reality technology and
aims at an improved excitement in welding, enabling the actual usage of gained theoretical know-how in a production
environment. Novel technologies in power supply technology, data communication and visualization improve the
user experience and user efficiency, assisted by simulations and optimization procedures, likewise helping the
training personnel to plan and guide learning and teaching experiences and interactive operation.
Fördergeber
- FRONIUS International GmbH
- Österreichische Forschungsförderungsgesellschaft mbH (FFG) , FFG
Externe Partner
- FH Campus Wien
- FH JOANNEUM Gesellschaft mbH
Beginn: 30.04.2017
MATAHARI - Maintenance through Assistive Telepresence And Human-centered Augmented Reality in Industry
Augmented Reality promises great advantages for empowering workers to perform increasingly complex
jobs in the era of industry 4.0. In this project, we focus on how Augmented Reality interfaces can be used
for tele-assistance, letting a remote expert support a worker in an industrial facility when it comes to
inspection, maintenance, construction and similar tasks. This encompasses efficient mobile 3D tracking
and scanning, annotating 3D scans with augmented information and transmitting the 3D data between
mobile workers and remote experts.
Fördergeber
- Österreichische Forschungsförderungsgesellschaft mbH (FFG) , FFG
Beginn: 31.03.2017
Fully Programmable GPU Pipelines
A modern computer system has a graphics processing unit (GPU) with enormous computational power. However, the overall hardware architecture of the GPU has not changed for almost 15 years. A fixed-function pipeline with a static sequence of stages supports a certain type of graphics application, primarily designed to deliver computer games. With new GPU programming languages such as CUDA, we can use the GPU power for computing simulations, but we cannot easily build new graphics pipelines, because certain parts of the GPU are not available using CUDA.
In this project, we propose to overcome this restriction with a new software framework, which supplements the missing parts of a graphics pipeline as a framework for a GPU programming language. This is challenging, because the framework has to be very efficient, or it will not be able to compete with a standard graphics pipeline. However, we can make use of the high flexibility afforded by a software implementation to produce a competitive framework.
More important than pure efficiency is that, with this software framework for graphics pipelines, we can then overcome all the restrictions of the conventional pipeline. For example, we can replace a rectangular dense framebuffer with a much more efficient irregular pixel structure. We can also replace uniform sampling patters for pixels with non-uniform ones, feeding into new Virtual Reality devices such as the Oculus Rift.
Fördergeber
- Fonds zur Förderung der wissenschaftlichen Forschung (FWF), FWF
Externe Partner
- Max-Planck-Institut für Informatik
Beginn: 28.02.2017
ARxX - Augmented Reality by Example
Augmented Reality (AR) by Example is a novel approach to generate AR applications from existing video and image demonstrations. Previous techniques create AR applications using a time-consuming process, which requires skills with 3D modeling and animation software, in addition to knowledge of the technical components of an AR system.
While any AR application will benefit from a simple authoring process, this project will mainly target the generation of AR applications for knowledge presentation in tutorials. While traditionally images have been used to provide visual instructions, with the success of video sharing platforms a large body of video tutorials is additionally available. Both, image and video tutorials effectively convey complex motions, but are difficult to follow precisely because of their 2D nature. AR tutorials have been revealed to be more effective.
This research projects brings the advantages of 2D and 3D instructions together by automatically creating three-dimensional AR tutorials from conventional 2D data. Unlike previous work, we will not simply overlay video, but synthesize 3D registered motion from the input data. Since the information in the resulting AR tutorial is registered to 3D objects, the user can freely change the viewpoint without degrading the AR experience.
This is achieved by investigating a number of different techniques. First, we have to extract the author's 3D environment and all 3D motion from the input data. Second, we need to provide tools for editing and retargeting the resulting 3D scene to the user's current environment. Third, we will develop comprehensible visualization techniques, specific for instructions in AR environments to effectively communicate the instructions extracted before. The research is complemented by qualitative and quantitative evaluations of the effects the investigated techniques have on users. This research project has great potential in applications concerned with crowed sourced training and teaching.
Fördergeber
- Fonds zur Förderung der wissenschaftlichen Forschung (FWF), FWF
Beginn: 28.02.2018
Augur - Augmented Reality for Measurement Instruments
Development Measuring distances
Fördergeber
- VRVis Zentrum für Virtual Reality und Visualisierung Forschungs-GmbH, VRVis
- Österreichische Forschungsförderungsgesellschaft mbH (FFG) , FFG
Beginn: 01.02.2015