We present a large-scale dataset of Planes in 3D, Pi3D, of roughly 1000 planes observed in 10 000 images from the 1DSfM dataset, and HEB, a large-scale homography estimation benchmark leveraging Pi3D. The applications of the Pi3D dataset are diverse, e.g. training or evaluating monocular depth, surface normal estimation and image matching algorithms. The HEB dataset consists of 226 260 homographies and includes roughly 4M correspondences. The homographies link images that often undergo significant viewpoint and illumination changes. As applications of HEB, we perform a rigorous evaluation of a wide range of robust estimators and deep learning-based correspondence filtering methods, establishing the current state-of-the-art in robust homography estimation. We also evaluate the uncertainty of the SIFT orientations and scales w.r.t. the ground truth coming from the underlying homographies and provide codes for comparing uncertainty of custom detectors. The dataset is available at https://github.com/danini/homography-benchmark.

in "A Realism Metric for Generated LiDAR Point Clouds" by Triess et al. In this work, a new metric is proposed that measures the quality of LiDAR point clouds that are generated, e.g., by a generative network. The metric can be used as an early indicator which assesses whether the generated training data will improve a down-stream task like semantic point cloud segmentation. Brissman et al. propose in "Recurrent Graph Neural Networks for Video Instance Segmentation" a very efficient approach for video instance segmentation. The approach tracks and segments multiple objects on-line and in real-time. Temporal information is also used in "Spatio-Temporal Outdoor Lighting Aggregation on Image Sequences using Transformer Networks" by Lee et al. to estimate outdoor illumination more consistently and without the need of an additional post-processing step.

Deriving semantic 3D models of man-made environments hitherto has not reached the desired maturity which makes human interaction obsolete. Man-made environments play a central role in navigation, city planning, building management systems, disaster management or augmented reality. They are characterised by rich geometric and semantic structures. These cause conceptual problems when learning generic models or when developing automatic acquisition systems. The problems appear to be caused by (1) the incoherence of the models for signal analysis, (2) the type of interplay between discrete and continuous geometric representations, (3) the inefficiency of the interaction between crisp models, such as partonomies and taxonomies, and soft models, mostly having a probabilistic nature, and (4) the vagueness of the used notions in the envisaged application domains. The paper wants to encourage the development and learning of generative models, specifically for man-made objects, to be able to understand, reason about, and explain interpretations.

Online pose estimation and mapping in unknown environments is essential for most mobile robots. Especially autonomous unmanned aerial vehicles require good pose estimates at comparably high frequencies. In this paper, we propose an effective system for online pose and simultaneous map estimation designed for light-weight UAVs. Our system consists of two components: (1) real-time pose estimation combining RTK-GPS and IMU at 100 Hz and (2) an effective SLAM solution running at 10 Hz using image data from an omnidirectional multi-fisheye-camera system. The SLAM procedure combines spatial resection computed based on the map that is incrementally refined through bundle adjustment and combines the image data with raw GPS observations and IMU data on keyframes. The overall system yields a real-time, georeferenced pose at 100 Hz in GPS-friendly situations. Additionally, we obtain a precise pose and feature map at 10 Hz even in cases where the GPS is not observable or underconstrained. Our system has been implemented and thoroughly tested on a 5 kg copter and yields accurate and reliable pose estimation at high frequencies. We compare the point cloud obtained by our method with a model generated from georeferenced terrestrial laser scanner.

The automatic extraction of line-networks from images is a well-known computer vision issue. Appearance and shape considerations have been deeply explored in the literature to improve accuracy in presence of occlusions, shadows, and a wide variety of irrelevant objects. However most existing works have ignored the structural aspect of the problem. We present an original method which provides structurally-coherent solutions. Contrary to the pixel-based and object-based methods, our result is a graph in which each node represents either a connection or an ending in the line-network. Based on stochastic geometry, we develop a new family of point processes consisting in sampling junction-points in the input image by using a Monte Carlo mechanism. The quality of a configuration is measured by a probability density which takes into account both image consistency and shape priors. Our experiments on a variety of problems illustrate the potential of our approach in terms of accuracy, flexibility and efficiency.

In this paper, we propose an incremental learning strategy for import vector machines (IVM), which is a sparse kernel logistic regression approach. We use the procedure for the concept of self-training for sequential classification of hyperspectral data. The strategy comprises the inclusion of new training samples to increase the classification accuracy and the deletion of noninformative samples to be memory and runtime efficient. Moreover, we update the parameters in the incremental IVM model without retraining from scratch. Therefore, the incremental classifier is able to deal with large data sets. The performance of the IVM in comparison to support vector machines (SVM) is evaluated in terms of accuracy, and experiments are conducted to assess the potential of the probabilistic outputs of the IVM. Experimental results demonstrate that the IVM and SVM perform similar in terms of classification accuracy. However, the number of import vectors is significantly lower when compared to the number of support vectors, and thus, the computation time during classification can be decreased. Moreover, the probabilities provided by IVM are more reliable, when compared to the probabilistic information, derived from an SVM's output. In addition, the proposed self-training strategy can increase the classification accuracy. Overall, the IVM and its incremental version is worthwhile for the classification of hyperspectral data.

A temporal filter approach for real-time detection and reconstruction of curbs and road surfaces from 3D point clouds is presented. Instead of local thresholding, as used in many other approaches, a 3D curb model is extracted from the point cloud. The 3D points are classified to different parts of the model (i.e. road and sidewalk) using a temporally integrated Conditional Random Field (CRF). The parameters of curb and road surface are then estimated from the respectively assigned points, providing a temporal connection via a Kalman filter.

This paper presents a generic framework for curb detection and reconstruction in the context of driver assistance systems. Based on a 3D point cloud, we estimate the parameters of a 3D curb model, incorporating also the curb adjacent surfaces, e.g. street and sidewalk. We apply an iterative two step approach. First, the measured 3D points, e.g., obtained from dense stereo vision, are assigned to the curb adjacent surfaces using loopy belief propagation on a Conditional Random Field. Based on this result, we reconstruct the surfaces and in particular the curb. Our system is not limited to straight-line curbs, i.e. it is able to deal with curbs of different curvature and varying height. The proposed algorithm runs in real-time on our demonstrator vehicle and is evaluated in urban real-world scenarios. It yields highly accurate results even for low curbs up to 20m distance.

We present a scheme for measuring completeness of local feature extraction in terms of image coding. Completeness is here considered as good coverage of relevant image information by the features. As each feature requires a certain number of bits which are representative for a certain subregion of the image, we interpret the coverage as a sparse coding scheme. The measure is therefore based on a comparison of two densities over the image domain: An entropy density pH(x) based on local image statistics, and a feature coding density pc(x) which is directly computed from each particular set of local features. Motivated by the coding scheme in JPEG, the entropy distribution is derived from the power spectrum of local patches around each pixel position in a statistically sound manner. As the total number of bits for coding the image and for representing it with local features may be different, we measure incompleteness by the Hellinger distance between pH(x) and pc(x). We will derive a procedure for measuring incompleteness of possibly mixed sets of local features and show results on standard datasets using some of the most popular region and keypoint detectors, including Lowe, MSER and the recently published SFOP detectors. Furthermore, we will draw some interesting conclusions about the complementarity of detectors.

This paper presents a novel method for detecting scale invariant keypoints. It fills a gap in the set of available methods, as it proposes a scale-selection mechanism for junction-type features. The method is a scale-space extension of the detector proposed by Förstner (1994) and uses the general spiral feature model of Bigün (1990) to unify different types of features within the same framework. By locally optimising the consistency of image regions with respect to the spiral model, we are able to detect and classify image structures with complementary properties over scale-space, especially star and circular shapes as interpretable and identifiable subclasses. Our motivation comes from calibrating images of structured scenes with poor texture, where blob detectors alone cannot find sufficiently many keypoints, while existing corner detectors fail due to the lack of scale invariance. The procedure can be controlled by semantically clear parameters. One obtains a set of keypoints with position, scale, type and consistency measure. We characterise the detector and show results on common benchmarks. It competes in repeatability with the Lowe detector, but finds more stable keypoints in poorly textured areas, and shows comparable or higher accuracy than other recent detectors. This makes it useful for both object recognition and camera calibration.

We present a geometric method for (i) matching 2D line segments from multiple oriented images, (ii) optimally reconstructing 3D line segments and (iii) grouping 3D line segments to corners. The proposed algorithm uses two developments in combining projective geometry and statistics, which are described in this article: (i) the geometric entities points, lines and planes in 2D and 3D and their uncertainty are represented in homogeneous coordinates and new entities may be constructed including their propagated uncertainty. The construction can be performed directly or as an estimation. (ii) relations such as incidence, equality, parallelism and orthogonality between points, lines and planes can be tested statistically based on a given significance level. Using these tools, the resulting algorithm is straightforward and gives reasonable results. It is only based on geometric information and does not use any image intensities, though it can be extended to use other information. The matching of 3D lines does not need any thresholds other than a significance value for the hypotheses tests.

The paper describes a new approach to image segmentation. It accepts the inherent deficiencies occuring when extracting low-level features and when dealing with the complexity of real scenes. Image segmentation therefore is understood as deriving a rich symbolic description useful for tasks such as stereo or object recognition in outdoor scenes. The approach is based on a polymorphic scheme for simultaneously extracting points, lines and segments in a topologically consistent manner, together with their mutual relations derived from the feature adjacency graph (FAG) thereby performing several grouping steps which gradually use more and more specific domain knowledge to achieve an optimal image description. The heart of the approach is (1) a detailed analysis of the FAG and (2) a robust estimation for validating the found geometric hypotheses. The analysis of the FAG, derived from the exoskeleton of the features, allows to detect inconsistencies of the extracted features with the ideal image model, a cell-complex. The FAG is used for finding hypotheses about incidence relations and geometric hypotheses, such as collinearity or parallelity, also between non-neighbored points and lines. The M-type robust estimation is used for simultaneously eliminating wrong hypotheses on geometric relationships. It uses a new argument for the weighting function.>

A minimum-description-length (MDL) principled evolutionary mechanism to automatic architecturing multilayer feedforward (MLFF) neural network is proposed. This type of neural networks is considered as a generic system implementing a generic model. The final network resultant from the architecturing and training is seen as an instance of this generic system, and thus an implemented instance of the generic model. Disregarding the hardware fault tolerance, the description length of the network and that of the performance deviation from the ideal given training samples must be smaller than that of the original samples. This total description length must be the minimum among all possible states. Constrained by the MDL principle, an MLFF neural network can be automatically architectured and trained through an evolutionary mechanism in which the network will be allowed and enabled to automatically expand as well as to reduce its complexity of architecture. The resultant network is of a partially connected MLFF architecture.>