A5063179993- Robotics and Sensor-Based Localization
- Advanced X-ray and CT Imaging
- Advanced Image and Video Retrieval Techniques
- Robot Manipulation and Learning
- Medical Imaging Techniques and Applications
- Geochemistry and Geologic Mapping
- Image Retrieval and Classification Techniques
- Robotic Mechanisms and Dynamics
- Manufacturing Process and Optimization
- Medical Imaging and Analysis
- Advanced Vision and Imaging
- Image Processing Techniques and Applications
- Visual Attention and Saliency Detection
- Human Pose and Action Recognition
- Optical measurement and interference techniques
- Radiomics and Machine Learning in Medical Imaging
- Multimodal Machine Learning Applications
- Advanced MIMO Systems Optimization
- Radiation Dose and Imaging
- Wireless Communication Networks Research
- Advanced Neural Network Applications
- Visual perception and processing mechanisms
- Soft Robotics and Applications
- Gaze Tracking and Assistive Technology
- Anomaly Detection Techniques and Applications
Technical University of Denmark
2017-2023
University of Southern Denmark
2013-2017
Maersk (Denmark)
2013-2016
Blekinge Institute of Technology
2010
Effective robot manipulation requires a vision system which can extract features of the environment determine what actions are possible. There is existing work in this direction under broad banner recognising "affordances". We particularly interested possibilities for afforded by relationships among pairs objects. For example if an object "inside" another or "on top" another. there need recognise such scene. use approach first segments image, and then considers pair objects to their physical...
We present an object recognition system coding shape by view-point invariant geometric relations and appearance. In our intelligent work-cell, the can observe work space of robot 3 pairs Kinect stereo cameras allowing for reliable complete information. show that in such a set-up we achieve high performance already with low number training samples. this to classify 56 objects using Random Forest algorithm. This indicates approach be used contexts as assembly manipulation which require...
The sensor-shifted stereo camera provides the mechanism for obtaining 3D information in a wide field of view.This novel kind requires simpler matching process comparison to convergence stereo.In addition this, uncertainty depth estimation target point space is defined by spatial quantization caused digital images.The dithering approach way reduce reconstruction through controlled adjustment parameters that shift levels.In this paper, mathematical model relates setup iso-disparities developed...
SUMMARY We present a fast object recognition system coding shape by viewpoint invariant geometric relations and appearance information. In our advanced industrial work-cell, the can observe work space of robot three pairs Kinect stereo cameras allowing for reliable complete From these sensors, we derive global features robust color making use normalization techniques. show that in such set-up, achieve high performance already with very low number training samples, which is crucial user...
This paper considers the spectrum sharing network consisting of a pair primary users (PUs) and cognitive (CRs) in fading channel. The PUs establishes wireless link as PU link. CRs CR utilize simultaneously with different priorities. has higher priority to respect When utilizes spectrum, desired quality service (QoS) is given be assured an opportunistic power scale under this constraint, assuring QoS on To compute optimal for link, fuzzy-based control strategy proposed based Mamdani fuzzy...
Recent developments in multispectral X-ray detectors allow for an efficient identification of materials based on their chemical composition. This has a range applications including security inspection, which is our motivation. In this paper, we analyze data from tomographic setup employing the MultiX detector, that records projection 128 energy bins covering 20 to 160 keV. Obtaining all information requires reconstructing tomograms, computationally expensive. Instead, propose reduce...
Spectral computed tomography (CT) is an emerging technology capable of providing high chemical specificity, which crucial for many applications such as detecting threats in luggage. This type application requires both fast and high-quality image reconstruction often based on sparse-view (few) projections. The conventional filtered back projection (FBP) method but it produces low-quality images dominated by noise artifacts CT. Iterative methods with, e.g., total variation regularizers can...
In this work, we learn a limited number of abstractions which can then be used to form preconditions for motor actions. These take the spatial relations amongst objects. We consider three "classes" relation: The objects either are separated from, on-top of, or inside each other. have tackled same problem in previous work (Fichtl et al., 2013). Here report on recent improved results using novel application histograms visually recognise relation between environment. Using histogram based...
In this paper, we present an object categorization system capable of assigning multiple and related categories for novel objects using multi-label learning. system, are described global geometric relations 3D features. We propose the Joint SVM method learning investigate extraction hierarchical clusters as a higher-level description to assist make comparisons with other approaches well single-label (including state-of-the-art methods different descriptors). The experiments carried out on...
Multi-spectral computed tomography is an emerging technology for the non-destructive identification of object materials and study their physical properties. Applications this can be found in various scientific industrial contexts, such as luggage scanning at airports. Material distinction its challenging, even with spectral x-ray information, due to acquisition noise, tomographic reconstruction artefacts setup application constraints. We present MUSIC - open access multi-spectral CT dataset...