Refine
Document Type
- Journal article (871)
- Conference proceeding (847)
- Book chapter (184)
- Book (61)
- Doctoral Thesis (34)
- Anthology (15)
- Working Paper (12)
- Patent / Standard / Guidelines (6)
- Review (6)
- Issue of a journal (2)
Language
- English (2041) (remove)
Is part of the Bibliography
- yes (2041)
Institute
- Informatik (702)
- ESB Business School (515)
- Technik (345)
- Life Sciences (327)
- Texoversum (150)
- Zentrale Einrichtungen (6)
Publisher
- Springer (299)
- IEEE (250)
- Elsevier (218)
- MDPI (98)
- Hochschule Reutlingen (55)
- Gesellschaft für Informatik (54)
- Wiley (49)
- ACM (40)
- De Gruyter (35)
- Association for Information Systems (AIS) (31)
For autonomously driving cars and intelligent vehicles it is crucial to understand the scene context including objects in the surrounding. A fundamental technique accomplishing this is scene labeling. That is, assigning a semantic class to each pixel in a scene image. This task is commonly tackled quite well by fully convolutional neural networks (FCN). Crucial factors are a small model size and a low execution time. This work presents the first method that exploits depth cues together with confidence estimates in a CNN. To this end, novel experimentally grounded network architecture is proposed to perform robust scene labeling that does not require costly preprocessing like CRFs or LSTMs as commonly used in related work. The effectiveness of this approach is demonstrated in an extensive evaluation on a challenging real-world dataset. The new architecture is highly optimized for high accuracy and low execution time.
This paper looks at the case of Reutlingen University (Hochschule Reutlingen), a university with a reputation for international student mobility. It examines how the university strives to fulfil its mandate to prepare ‘industry-ready’ graduates for the global industry by providing an international practice-oriented education. The key focus is on its efforts to establish credit-bearing internship programmes for international students, an area where the institution has ramped up its activities in recent years. Internships for international students is understood to encompass both domestic internships for international students (exchange and degree-seeking students) as well as internships abroad for home-grown degree-seeking students. The paper presents models and approaches that seek to ensure the quality of the international internship experience. It discusses challenges that the university has encountered on the way and makes suggestions about how to create internship opportunities against the backdrop of competing demands and expectations.
Turbidity sensing is very common in the control of drinking water. Furthermore, turbidity measurements are applied in the chemical (e.g., process monitoring), pharmaceutical (e.g., drug discovery), and food industries (e.g., the filtration of wine and beer). The most common measurement technique is nephelometric turbidimetry. A nephelometer is a device for measuring the amount of scattered light of suspended particles in a liquid by using a light source and a light detector orientated in 90°to each other. Commercially available nephelometers cost usually—depending on the measurable range, reliability, and precision —thousands of euros. In contrast, our new developed GRIN-lens-based nephelometer, called GRINephy, combines low costs with excellent reproducibility and precision, even at very low turbidity levels, which is achieved by its ability to rotate the sample. Thereby, many cuvette positions can be measured, which results in a more precise average value for the turbidity calculated by an algorithm, which also eliminates errors caused by scratches and contaminations on the cuvettes. With our compact and cheap Arduino-based sensor, we are able to measure in the range of 0.1–1000 NTU and confirm the ISO 7027-1:2016 for low turbidity values.
There are several intra-operative use cases which require the surgeon to interact with medical devices. I used the Leap Motion Controller as input device for three use-cases: 2D-interaction (e.g. advancing EPR data), selection of a value (e.g. room illumination brightness) and an application point and click scenario. I evaluated the Palm Mouse as the most suitable gesture solution to coordinate the mouse and advise to use the implementation using all fingers to perform a click. This small case study introduces the implementations and methods that result those recommendations.
Reconstructing 3D face shape from a single 2D photograph as well as from video is an inherently ill-posed problem with many ambiguities. One way to solve some of the ambiguities is using a 3D face model to aid the task. 3D morphable face models (3DMMs) are amongst the state of the art methods for 3D face reconstruction, or so called 3D model fitting. However, current existing methods have severe limitations, and most of them have not been trialled on in-the-wild data. Current analysis-by- synthesis methods form complex non linear optimisation processes, and optimisers often get stuck in local optima. Further, most existing methods are slow, requiring in the order of minutes to process one photograph.
This thesis presents an algorithm to reconstruct 3D face shape from a single image as well as from sets of images or video frames in real-time. We introduce a solution for linear fitting of a PCA shape identity model and expression blendshapes to 2D facial landmarks. To improve the accuracy of the shape, a fast face contour fitting algorithm is introduced. These different components of the algorithm are run in iteration, resulting in a fast, linear shape-to- landmarks fitting algorithm. The algorithm, specifically designed to fit to landmarks obtained from in-the-wild images, by tackling imaging conditions that occur in in-the-wild images like facial expressions and the mismatch of 2D–3D contour correspondences, achieves the shape reconstruction accuracy of much more complex, nonlinear state of the art methods, while being multiple orders of magnitudes faster.
Second, we address the problem of fitting to sets of multiple images of the same person, as well as monocular video sequences. We extend the proposed shape-to-landmarks fitting to multiple frames by using the knowledge that all images are from the same identity. To recover facial texture, the approach uses texture from the original images, instead of employing the often-used PCA albedo model of a 3DMM. We employ an algorithm that merges texture from multiple frames in real-time based on a weighting of each triangle of the reconstructed shape mesh.
Last, we make the proposed real-time 3D morphable face model fitting algorithm available as open-source software. In contrast to ubiquitous available 2D-based face models and code, there is a general lack of software for 3D morphable face model fitting, hindering a widespread adoption. The library thus constitutes a significant contribution to the community.
Thematic issue on human-centred ambient intelligence: cognitive approaches, reasoning and learning
(2017)
This editorial presents advances on human-centred Ambient Intelligence applications which take into account cognitive issues when modelling users (i.e. stress, attention disorders), and learn users’ activities/preferences and adapt to them (i.e. at home, driving a car). These papers also show AmI applications in health and education, which make them even more valuable for the general society.
Coupling electricity and heat sector is one of the most necessary actions for the successful energy transition. Efficient electrification for space heating and domestic hot water generation is needed for buildings, which are not connected to any district heating network, as distributed heating demand momentarily is largely met by fossil fuels. Hence, hybrid energy systems will play a pivotal role for the energy transition in buildings. Heat pumps running on PV-electricity is one of the most widely discussed combination for this purpose. In this paper, a heuristic optimization method for the optimal operation of a heat pump driven by the objective for maximum onsite PV electricity utilization is presented. In this context, the thermal flexibility of the building and a thermal energy storage (TES) for generation of domestic hot water (DHW) are activated in order to shift the operation of the heat pump to times of PV-generation. Yearly simulations for a system consisting of heat pump, PV modules, building with floor heating installation and TES for DHW generation are carried out. Variation parameters for the simulation include room temperature amplitude (0.5, 1, 1.5 and 2 K) based on mean room temperature (21 °C), PV-capacity (4, 6, 8 and 10 kW) and type of heat pump (ground source and air source type). The yearly energy balances show that buildings offer significant thermal storage capacity avoiding an additional, large TES for space heating fulfillment and improving the share of onsite PV electricity utilization. With introduction of a battery, which has been analyzed as well for different sizes (1.9, 4.8, 7.7 and 10.6 kWh), the share of onsite PVelectricity utilization can even be improved. However, thermal flexibility supplemented by the varying room temperature amplitude for a bigger battery does not improve the share of onsite PV-electricity utilization. Nevertheless, even with a battery not more than 50% of the electrical load including operation of the heat pump can be covered by PV-electricity for the specific system under investigation. This is noteworthy on the one hand, since it indicates that a hybrid heating system consisting of heat pump and PV cannot solely cover the heat demand of residential buildings. One the other hand, this emphasizes the necessity to include further renewable sources like wind power, in order to draw the complete picture. This, however, is beyond the scope of this paper, which mainly focuses on introduction and verification of the novel control method with regard to a practical building.
The main challenge when driving heat pumps by PV-electricity is balancing differing electrical and thermal demands. In this article, a heuristic method for optimal operation of a heat pump driven by a maximum share of PV-electricity is presented. For this purpose, the (DHW) are activated in order shift the operation of the heat pump to times of PV-generation. The system under consideration refers to thermal and electrical demands of a single family house. It consists of a heat pump, a thermal energy storage for DHW and of grid connected heating and generation of domestic hot water, the heat pump runs with two different supply temperatures and thereby achieving a maximum overall COP. Within the algorithm for optimization a set of heuristic rules is developed in a way that the operational characteristics of the heat pump in terms of minimum running and stopping times are met as well as the limiting constraints of upper and lower limits of room temperature and energy content of electricity generated, a varying number of heat pump schedules fulfilling the bundary conditions are created. Finally, the schedule offering the maximum on-site utilization of PV-electricity with a minimum number of starts of the heat pump, which serves as secondary condition, is selected. Yearly simulations of this combination have been carried out. Initial results of this method indicate a significant rise in on-site consumption of the PV-electricity and heating demand fulfilment by renewable electricity with no need for a massive TES for the heating system in terms of a big water tank.
Consistent supply chain management across all levels of value creation is a common approach in the industrial sector. The implementation in agricultural processes requires rethinking in the supply chain concept. The reasons are the heuristic characterized processes, the stochastic environmental conditions, the mobility of the production facilities and the low division of work.
In this paper we deal with how concepts of innovative supply chain management of Industrie 4.0 could not only deliver a way to overcome said problems but also provide the foundation for the development of new forms of work and business models for Farming 4.0.