A5050558405- Medical Image Segmentation Techniques
- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Industrial Vision Systems and Defect Detection
- Autonomous Vehicle Technology and Safety
- Image and Object Detection Techniques
- Medical Imaging and Analysis
- Advanced X-ray and CT Imaging
- Radiomics and Machine Learning in Medical Imaging
- Advanced Neural Network Applications
- Medical Imaging Techniques and Applications
- Advanced Image and Video Retrieval Techniques
- Image and Signal Denoising Methods
- Reinforcement Learning in Robotics
- Image Processing Techniques and Applications
- Video Surveillance and Tracking Methods
- Social Robot Interaction and HRI
- Image Enhancement Techniques
- Anomaly Detection Techniques and Applications
- Face and Expression Recognition
- Neural Networks and Applications
- Robotics and Automated Systems
- 3D Surveying and Cultural Heritage
- Advanced Image Fusion Techniques
- Aortic aneurysm repair treatments
University of Duisburg-Essen
2015-2024
Thermo Fisher Scientific (Japan)
2024
Federal Institute For Materials Research and Testing
2021
Technical University of Munich
1988-2005
Kiel University
1997-2002
Fraunhofer Society
2002
In this paper a novel spline-based multi-lane detection and tracking system is proposed. Reliable lane an important component of departure warning systems, keeping support systems or change assistance systems. The major novelty the proposed approach usage so-called Catmull-Rom spline in combination with extended Kalman filter tracking. new model enables accurate flexible modeling markings. At same time application contributes significantly to robustness stability. There no assumption about...
This paper presents a novel approach to obtain reliable and robust time-to-contact estimates from monocular moving camera observing various obstacles. The algorithm utilizes interest points measure the relative scale change of an obstacle applies estimation techniques combine different measurements into one three possible motion models. These include model with constant distance, velocity acceleration. An interacting multiple framework is used select appropriate finally estimate observed...
This paper presents an automotive video and radar data fusion framework that can be used as a preliminary stage of automatic cruise control or collision mitigation by braking system. The finds the optimal assignment camera target reports provides improved state estimates for fused targets. A sophisticated critical path selection is presented in module aims to select most relevant target. capable identifying targets cut into ego lane out from incorporate final selection. selected then...
Nowadays, the camera online calibration module has been a fundamental and often requested component in an Advanced Driver Assistance System (ADAS). The proposed system provides efficient practical solution to such request, also shows technical advances compared earlier systems. It utilizes lane markings for orientation calibration. At each image frame, multiple vanishing points of are estimated using weighted least squares method, followed by tracking process with Kalman Filter better...
Abstract Background Since 2001, Serum Free Light Chain measurements have become established as routine clinical laboratory tests. Katzmann (2002) published a reference interval based on 282 normal samples establishing ratio range 0.26-1.65. Whilst kappa and lambda are of little diagnostic value, kappa/lambda gives strong indication monoclonal gammopathy. In recent years, it has been suggested that the may changed. Here we present investigations identifying protein/protein interactions...
Measuring absolute distances in monocular vision is challenging and cannot be solved directly. Conventionally, assumptions like an a priori width of the target object or geometric constraints are made to overcome problem.
We propose a 2D Encoder-Decoder based deep learning architecture for semantic segmentation, that incorporates anatomical priors by imitating the encoder component of an autoencoder in latent space. The is additionally enhanced means hierarchical features, extracted UNet module. Our suggested trained end-to-end manner and evaluated on example pelvic bone segmentation MRI. A comparison to standard U-Net shows promising improvements.
With an application to the quality control of steel, we present image processing algorithms for unsupervised detection anomalies that are hidden within a global milling pattern. Thereby, consider Fourier-based approaches and localized shearlet decomposition damping texture. These frequency-based compared template-based autocorrelation in spatial domain. As alternative approach, so-called matching pursuit is proposed with joint Fourier wavelet-based dictionary, whereby two models grooves can...
Inference and reasoning in modern AmI (Ambient Intelligence) middlewares is still a complex task. Currently no common patterns for building smart applications can be identified. This paper presents an ongoing effort to build generic probabilistic framework the networked homes. The utilized designing agents systematic unified way. developed modeling algorithms make extensive use of information about user way he/she interacts with system. To achieve this, several levels knowledge...