Reinhard Klette

Differences between stereo and motion behaviour on synthetic and real-world stereo sequences

Performance evaluation of stereo or motion analysis techniques is commonly done either on synthetic data where the ground truth can be calculated from ray-tracing principals, or on engineered data where ground truth is easy to estimate. Furthermore, these scenes are usually only shown in a very short sequence of images. This paper shows why synthetic scenes may not be the only testing criteria by giving evidence of conflicting results of disparity and optical flow estimation for real-world and synthetic testing. The data dealt with in this paper are images taken from a moving vehicle. Each real-world sequence contains 250 image pairs or more. Synthetic driver assistance scenes (with ground truth) are 100 or more image pairs. Particular emphasis is paid to the estimation and evaluation of scene flow on the synthetic stereo sequences. All image data used in this paper is made publicly available at http: //www.mi.auckland.ac.nz/EISATS.

Research in the theory of inductive inference by GDR mathematicians-A survey

Recent results in the theory of inductive inference are summarized. They concern deciphering of automata, language identification, prediction of functions, inference with additional information, strategies, functionals, index sets, characterization of identification types, uniform inference, and inference of nonrandom sequences. For proofs and further results in the field of inductive inference due to mathematicians of the German Democratic Republic a detailed bibliography is included.

A comparative evaluation of length estimators of digital curves

This paper compares previously published length estimators in image analysis having digitized curves as input. The evaluation uses multigrid convergence (theoretical results and measured speed of convergence) and further measures as criteria. This paper also suggests a new gradient-based method for length estimation, and combines a previously proposed length estimator for straight segments with a polygonalization method.

Digital planarity-A review

Digital planarity is defined by digitizing Euclidean planes in the three-dimensional digital space of voxels; voxels are given either in the grid-point or the grid-cube model. The paper summarizes results (also including most of the proofs) about different aspects of digital planarity, such as supporting or separating Euclidean planes, characterizations in arithmetic geometry, periodicity, connectivity, and algorithmic solutions. The paper provides a uniform presentation, which further extends and details a recent book chapter in [R. Klette, A. Rosenfeld, Digital Geometry-Geometric Methods for Digital Picture Analysis, Morgan Kaufmann, San Francisco, 2004].

Robust background subtraction and maintenance

Background subtraction is one of the main techniques to extract moving objects from background scenes. A mixture of Gaussians is a common model for background subtraction that has been used in many applications. However modelling background pixels using this model results into a low-level process at pixel level. Some of its main drawbacks are: a subtracted (moving object) region may contain holes; it cannot solve partial occlusion problems, and it requires updates in cases of shadows or sudden changes in the scene. We present a multi-layered mixture of Gaussians model named PixelMap. We combine the mixture of Gaussians model with concepts defined by region level and frame level considerations. Our experimental results show that our method improved the accuracy of extracting moving objects from background. A single stationary camera has been used.

Iterative semi-global matching for robust driver assistance systems

Semi-global matching (SGM) is a technique of choice for dense stereo estimation in current industrial driver-assistance systems due to its real-time processing capability and its convincing performance. In this paper we introduce iSGM as a new cost integration concept for semi-global matching. In iSGM, accumulated costs are iteratively evaluated and intermediate disparity results serve as input to generate semi-global distance maps. This novel data structure supports fast analysis of spatial disparity information and allows for reliable search space reduction in consecutive cost accumulation. As a consequence horizontal costs are stabilized which improves the robustness of the matching result. We demonstrate the superiority of this iterative integration concept against a standard configuration of semi-global matching and compare our results to current state-of-the-art methods on the KITTI Vision Benchmark Suite.

Human Motion – Understanding, Modeling, Capture and Animation

Motion Capture and Pose Estimation.- Marker-Less 3D Feature Tracking for Mesh-Based Human Motion Capture.- Boosted Multiple Deformable Trees for Parsing Human Poses.- Gradient-Enhanced Particle Filter for Vision-Based Motion Capture.- Multi-activity Tracking in LLE Body Pose Space.- Exploiting Spatio-temporal Constraints for Robust 2D Pose Tracking.- Efficient Upper Body Pose Estimation from a Single Image or a Sequence.- Real-Time and Markerless 3D Human Motion Capture Using Multiple Views.- Modeling Human Locomotion with Topologically Constrained Latent Variable Models.- Silhouette Based Generic Model Adaptation for Marker-Less Motion Capturing.- Body and Limb Tracking and Segmentation.- 3D Hand Tracking in a Stochastic Approximation Setting.- Nonparametric Density Estimation with Adaptive, Anisotropic Kernels for Human Motion Tracking.- Multi Person Tracking Within Crowded Scenes.- Joint Appearance and Deformable Shape for Nonparametric Segmentation.- Robust Spectral 3D-Bodypart Segmentation Along Time.- Articulated Object Registration Using Simulated Physical Force/Moment for 3D Human Motion Tracking.- An Ease-of-Use Stereo-Based Particle Filter for Tracking Under Occlusion.- Activity Recognition.- Semi-Latent Dirichlet Allocation: A Hierarchical Model for Human Action Recognition.- Recognizing Activities with Multiple Cues.- Human Action Recognition Using Distribution of Oriented Rectangular Patches.- Human Motion Recognition Using Isomap and Dynamic Time Warping.- Behavior Histograms for Action Recognition and Human Detection.- Learning Actions Using Robust String Kernels.

Multigrid Convergence of Calculated Features in Image Analysis

This paper informs about number-theoretical and geometrical estimates of worst-case bounds for quantization errors in calculating features such as moments, moment based features, or perimeters in image analysis, and about probability-theoretical estimates of error bounds (e.g. standard deviations) for such digital approximations. New estimates (with proofs) and a review of previously known results are provided.

Quantitative color optical flow

We perform a qualitative and quantitative analysis of various multi-frame color optical flow methods for synthetic and real panning and zooming image sequences. We show that optical flow accuracy improvement can be slightly improved if color images are available instead of gray value or saturation images. We show the usefulness of a directional regularization constraint for computing optical flow when the camera motion is known to be panning or zooming.

Image Stitching - Comparisons and New Techniques

This work describes the steps involved in the generation of a panoramic image, i.e., image acquisition, image registration and image merging. Different approaches for each of these steps are discussed and compared. The resultant images from image registration and image merging are quantitatively evaluated to provide an indication to the performance of the different methods. The methods provided in this work can be used to generate panoramic images for the use of interactive panoramic viewing of images, architectural walk-through, and other applications associated with the modelling of 3D environments.