A5011359067- Advanced Image and Video Retrieval Techniques
- Advanced Vision and Imaging
- Robotics and Sensor-Based Localization
- Video Surveillance and Tracking Methods
- Medical Image Segmentation Techniques
- Image Retrieval and Classification Techniques
- 3D Surveying and Cultural Heritage
- Face and Expression Recognition
- Human Pose and Action Recognition
- Advanced Neural Network Applications
- Anomaly Detection Techniques and Applications
- Neural Networks and Applications
- Face recognition and analysis
- Remote Sensing and LiDAR Applications
- Domain Adaptation and Few-Shot Learning
- Image Processing Techniques and Applications
- Image and Object Detection Techniques
- Optical measurement and interference techniques
- Industrial Vision Systems and Defect Detection
- Advanced Image Processing Techniques
- 3D Shape Modeling and Analysis
- Image Processing and 3D Reconstruction
- Remote-Sensing Image Classification
- Computer Graphics and Visualization Techniques
- Medical Imaging and Analysis
Graz University of Technology
2016-2025
Christian Doppler Laboratory for Thermoelectricity
2021-2024
DIPF | Leibniz Institute for Research and Information in Education
2024
Institute of Computer Vision and Applied Computer Sciences
2008-2019
University of Ulster
2019
Toyota Technological Institute
2019
KTH Royal Institute of Technology
2019
Keio University
2019
University of Bern
2019
BioTechMed-Graz
2014-2018
The Photodetector Array Camera and Spectrometer (PACS) is one of the three science instruments on ESA's far infrared submillimetre observatory. It employs two Ge:Ga photoconductor arrays (stressed unstressed) with 16x25 pixels, each, filled silicon bolometer 16x32 32x64 respectively, to perform integral-field spectroscopy imaging photometry in 60-210μ m wavelength regime. In mode, it simultaneously images bands, 60-85μ or 85-125μ\m 125-210μ m, over a field view ~1.75'x3.5', close Nyquist...
In this paper, we raise important issues on scalability and the required degree of supervision existing Mahalanobis metric learning methods. Often rather tedious optimization procedures are applied that become computationally intractable a large scale. Further, if one considers constantly growing amount data it is often infeasible to specify fully supervised labels for all points. Instead, easier in form equivalence constraints. We introduce simple though effective strategy learn distance...
<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> This paper presents a comparison study between 10 automatic and six interactive methods for liver segmentation from contrast-enhanced CT images. It is based on results the "MICCAI 2007 Grand Challenge" workshop, where 16 teams evaluated their algorithms common database. A collection of 20 clinical images with reference segmentations was provided to train tune in advance. Participants were also...
Boosting has become very popular in computer vision, showing impressive performance detection and recognition tasks. Mainly off-line training methods have been used, which implies that all data to be a priori given; usage of the classifier are separate steps. Training on-line incrementally as new becomes available several advantages opens areas application for boosting vision. In this paper we propose novel AdaBoost feature selection method. conjunction with efficient extraction method is...
Face alignment is a crucial step in face recognition tasks. Especially, using landmark localization for geometric normalization has shown to be very effective, clearly improving the results. However, no adequate databases exist that provide sufficient number of annotated facial landmarks. The are either limited frontal views, only small images or have been acquired under controlled conditions. Hence, we introduce novel database overcoming these limitations: Annotated Facial Landmarks Wild...
The aim of single image super-resolution is to reconstruct a high-resolution from low-resolution input. Although the task ill-posed it can be seen as finding non-linear mapping low high-dimensional space. Recent methods that rely on both neighborhood embedding and sparse-coding have led tremendous quality improvements. Yet, many previous approaches are hard apply in practice because they either too slow or demand tedious parameter tweaks. In this paper, we propose directly map patches using...
The Visual Object Tracking challenge 2015, VOT2015, aims at comparing short-term single-object visual trackers that do not apply pre-learned models of object appearance. Results 62 are presented. number tested makes VOT 2015 the largest benchmark on tracking to date. For each participating tracker, a short description is provided in appendix. Features VOT2015 go beyond its VOT2014 predecessor are: (i) new dataset twice as large with full annotation targets by rotated bounding boxes and...
In this work we present a novel method for the challenging problem of depth image up sampling. Modern cameras such as Kinect or Time-of-Flight deliver dense, high quality measurements but are limited in their lateral resolution. To overcome limitation formulate convex optimization using higher order regularization an isotropic diffusion tensor, calculated from resolution intensity image, is used to guide We derive numerical algorithm based on primal-dual formulation that efficiently...
Efficient view registration with respect to a given 3D reconstruction has many applications like inside-out tracking in indoor and outdoor environments, geo-locating images from large photo collections. We present fast location recognition technique based on structure motion point clouds. Vocabulary tree-based indexing of features directly returns relevant fragments models instead documents the database. Additionally, we propose compressed scene representation which improves rates while...
Random Forests (RFs) are frequently used in many computer vision and machine learning applications. Their popularity is mainly driven by their high computational efficiency during both training evaluation while achieving state-of-the-art results. However, most applications RFs off-line. This limits usability for practical problems, instance, when data arrives sequentially or the underlying distribution continuously changing. In this paper, we propose a novel on-line random forest algorithm....
The step away from a synchronized or cue-based brain-computer interface (BCI) and laboratory conditions towards real world applications is very important crucial in BCI research. This work shows that ten naive subjects can be trained synchronous paradigm within three sessions to navigate freely through virtual apartment, whereby at every junction the could decide by their own, how they wanted explore environment (VE). apartment was designed similar application, with goal-oriented task, high...
The authors report the application of three-layer back-propagation networks for classification Landsat TM data on a pixel-by-pixel basis. results are compared to Gaussian maximum likelihood classification. First, it is shown that neural network able perform better than classifier. Secondly, in an extension basic architecture textural information can be integrated into classifier without explicit definition texture measure. Finally, use postclassification smoothing examined.< <ETX...
To determine and compare the performance of different classifiers applied to four-class EEG data is goal this communication. The were recorded with 60 electrodes from five subjects performing four motor-imagery tasks. signal was modeled by an adaptive autoregressive (AAR) process whose parameters extracted Kalman filtering. By these AAR obtained, namely minimum distance analysis (MDA)—for single-channel analysis, linear discriminant (LDA), k-nearest-neighbor (kNN) as well support vector...
Tracking-by-detection is increasingly popular in order to tackle the visual tracking problem. Existing adaptive methods suffer from drifting problem, since they rely on self-updates of an on-line learning method. In contrast previous work that tackled this problem by employing semi-supervised or multiple-instance learning, we show augmenting method with complementary approaches can lead more stable results. particular, use a simple template model as non-adaptive and thus component, novel...
In this work we revisit the Mumford-Shah functional, one of most studied variational approaches to image segmentation. The contribution paper is propose an algorithm which allows minimize a convex relaxation functional obtained by lifting. efficient primal-dual projection for prove convergence. contrast existing algorithms minimizing full first based on relaxation. As consequence computed solutions are independent initialization. Experimental results confirm that proposed determines smooth...
In this paper, we address the problem of model-free online object tracking based on color representations. According to findings recent benchmark evaluations, such trackers often tend drift towards regions which exhibit a similar appearance compared interest. To overcome limitation, propose an efficient discriminative model allows us identify potentially distracting in advance. Furthermore, exploit knowledge adapt representation beforehand so that distractors are suppressed and risk drifting...