MI-tech Reports (2007+)

Permanent URI for this collectionhttps://hdl.handle.net/2292/3205

Browse

Now showing 1 - 20 of 47

Discrete Driver Assistance
(Computer Science Department, The University of Auckland, New Zealand, 2009) Klette, Reinhard; Jiang, Ruyi; Morales, Sandino; Vaudrey, Tobi
Applying computer technology, such as computer vision in driver assistance, implies that processes and data are modeled as being discretized rather than being continuous. The area of stereo vision provides various examples how concepts known in discrete mathematics (e.g., pixel adjacency graphs, belief propagation, dynamic programming, max- ow/min-cut, or digital straight lines) are applied when aiming for efficient and accurate pixel correspondence solutions. The paper reviews such developments for a reader in discrete mathematics who is interested in applied research (in particular, in vision-based driver assistance). As a second subject, the paper also discusses lane detection and tracking, which is a particular task in driver assistance; recently the Euclidean distance transform proved to be a very appropriate tool for obtaining a fairly robust solution.
Improving Optical Flow using Residual Images
(Computer Science Department, The University of Auckland, New Zealand, 2009) Vaudrey, Tobi; Wedel, Andreas; Klette, Reinhard
Optical flow is a highly researched area in low-level computer vision. It is a complex problem which tries to solve a 2D search in continuous space, while the input data is 2D discrete data. The major assumption in most optical flow applications is the intensity consistency assumption, introduced by Horn and Schunck. This constraint is often violated in practice. This paper proposes and generalises one such approach; using residual images (high-frequencies) of images, to remove the illumination differences between corresponding images.
Fractal Art based on "The Butterfly Effect" of Chaos Theory
(Computer Science Department, The University of Auckland, New Zealand, 2009) Chang, Yin-Wei; Huang, Fay
This paper proposes and demonstrates a new integration of the theory of fractals and the butterfly effect of chaos theory. They both have long histories in creating digital artworks, but besides of many existing fractal software programs, none of them allowed us to achieve the proposed integration. Moreover, our program is the first to provide the functional concepts of overlapping results and sequential transformations, which allow us to generate a wider variety of patterns. Our program not only has the potential of creating 2D digital artworks but also supports the creation of animated abstract artworks.
Tracking small artists
(Computer Science Department, The University of Auckland, New Zealand, 2009) Russell, James; Klette, Reinhard
Tracks of small animals are important in environmental surveillance, where pattern recognition algorithms allow species identification of the individuals creating tracks. These individuals can also be seen as artists, presented in their natural environments with a canvas upon which they can make prints. We present tracks of small mammals and reptiles which have been collected for identification purposes, and reinterpret them from an esthetic point of view. We re-classify these tracks not by their geometric qualities as pattern recognition algorithms would, but through interpreting the `artist', their brush strokes and intensity. We describe the algorithms used to enhance and present the work of the `artists'.
Lane Detection on the iPhone
(Computer Science Department, The University of Auckland, New Zealand, 2009) Ren, Feixiang; Huang, Jinsheng; Jiang, Ruyi; Klette, Reinhard
A robust and efficient lane detection system is an essential component of Lane Departure Warning Systems, which are commonly used in many vision-based Driver Assistance Systems (DAS) in intelligent transportation. Various computation platforms have been proposed in the past few years for the implementation of driver assistance systems (e.g., PC, laptop, integrated chips, play station, and so on). In this paper, we propose a new platform for the implementation of lane detection, which is based on a mobile phone (the iPhone). Due to physical limitations of the iPhone w.r.t. memory and computing power, a simple and efficient lane detection algorithm using a Hough transform is developed and implemented on the iPhone, as existing algorithms developed based on the PC platform are not suitable for mobile phone devices (currently). Experiments of the lane detection algorithm are made both on PC and on iPhone.
Low-level Image Processing for Lane Detection and Tracking
(Computer Science Department, The University of Auckland, New Zealand, 2009) Jiang, Ruyi; Klette, Reinhard; Wang, Shigang; Vaudrey, Tobi
Lane detection and tracking is a significant component of vision-based driver assistance systems (DAS). Low-level image processing is the first step in such a component. This paper suggests three useful techniques for low-level image processing in lane detection situations: bird’s-eye view mapping, a specialized edge detection method, and the distance transform. The first two techniques have been widely used in DAS, while the distance transform is a method newly exploited in DAS, that can provide useful information in lane detection situations. This paper recalls two methods to generate a bird’s-eye image from the original input image, it also compares edge detectors. A modified version of the Euclidean distance transform called real orientation distance transform (RODT) is proposed. Finally, the paper discusses experiments on lane detection and tracking using these technologies.
A Methodology for Evaluating Illumination Artifact Removal for Corresponding Images
(Computer Science Department, The University of Auckland, New Zealand, 2009) Vaudrey, Tobi; Wedel, Andreas; Klette, Reinhard
Robust stereo and optical flow disparity matching is essential for computer vision applications with varying illumination conditions. Most robust disparity matching algorithms rely on computationally expensive normalized variants of the brightness constancy assumption to compute the matching criterion. In this paper, we reinvestigate the removal of global and large area illumination artifacts, such as vignetting, camera gain, and shading reflections, by directly modifying the input images. We show that this significantly reduces violations of the brightness constancy assumption, while maintaining the information content in the images. In particular, we define metrics and perform a methodical evaluation to firstly identify the loss of information in the images, and secondly determine the reduction of brightness constancy violations. Thirdly, we experimentally validate that modifying the input images yields robustness against illumination artifacts for optical flow disparity matching.
Fast Trilateral Filtering
(Computer Science Department, The University of Auckland, New Zealand, 2009) Vaudrey, Tobi; Klette, Reinhard
This paper compares the original implementation of the trilateral filter with two proposed speed improvements. One is using simple look-up-tables, and leads to exactly the same results as the original filter. The other technique is using a novel way of truncating the look-up-table to a user specified required accuracy. Here, results differ from those of the original filter, but to a very minor extent. The paper shows that measured speed improvements of this second technique are in the order of several magnitudes, compared to the original or LUT trilateral filter.
Lane Detection and Tracking Using a New Lane Model and a Distance Transform
(Computer Science Department, The University of Auckland, New Zealand, 2009) Jiang, Ruyi; Klette, Reinhard; Wang, Shigang; Vaudrey, Tobi
Lane detection is an important component of driver assistance systems (DAS), and highway-based lane departure solutions have been in the market since the mid 1990s. However, improving and generalizing vision-based lane detection solutions remains to be a challenging task. Particle filtering of boundary points is a robust way to estimate lanes. This paper introduces a new lane model in correspondence to this particle filter-based approach. Furthermore, a modified version of an Euclidean distance transform is applied on an edge map to provide information for boundary point detection. In comparison to the edge map, properties of the distance transform support improved lane detection including a novel initialization method. Two lane tracking methods are also discussed while focusing on efficiency and robustness, respectively. Finally, the paper reports about experiments on lane detection and tracking.
Prediction Error Evaluation
(Computer Science Department, The University of Auckland, New Zealand, 2009) Morales, Sandino; Klette, Reinhard
Prediction errors are commonly used when analyzing the performance of a multi-camera stereo system using at least three cameras. This paper discusses this methodology for performance evaluation on long stereo sequences (in the context of vision-based driver assistance systems). Three cameras are calibrated in an ego-vehicle, and prediction error analysis is performed on recorded stereo sequences. They are evaluated using various common stereo matching algorithms, such as belief propagation, dynamic programming, semi-global matching, or graph cut. This performance evaluation is demonstrated on synthetic and real data.
An Approximate Algorithm for Solving Shortest Path Problems for Mobile Robots or Driver Assistance
(Computer Science Department, The University of Auckland, New Zealand, 2009) Li, Fajie; Klette, Reinhard; Morales, Sandino
Finding a shortest path between two given locations is of importance for mobile robots, but also (e.g.) for identifying unique paths in a given surrounding region II when (e.g.) evaluating vision software in test vehicles, or for calculating the free-space boundary in vision-based driver assistance. We assume that II is given as a triangulated surface which is not necessary simply connected.
Graph-Cut and Belief-Propagation Stereo
(Computer Science Department, The University of Auckland, New Zealand, 2009) Joachim, Penc; Reinhard, Klette; Tobi, Vaudrey; Sandino, Morales
This paper deals with stereo correspondence search, using graph cuts and belief propagation, for estimating depth maps. The results following different preprocessing steps are evaluated, using the quality of the disparity map. Running times are also investigated. For evaluation purposes, different kinds of images have been used: reference images from the Middlebury Stereo website, synthetic driving scenes, and real-world stereo sequences, either provided by Daimler AG or self-recorded with the research vehicle of the .enpeda.. project at Auckland University.
Residual Images Remove
(Computer Science Department, The University of Auckland, New Zealand, 2009) Vaudrey, Tobi; Reinhard, Klette
Real-world image sequences (e.g., recorded for vision-based driver assistance) are typically degraded by various types of noise, changes in lighting, out-of-focus lenses, differing exposures, and so forth. In past studies, illumination effects have been proven to cause the most common problems in correspondence algorithms. We address this problem using the concept of residuals, which is the difference between an image and a smoothed version of itself. In this paper, we conduct a study identifying that the residual images contain the important information in an image. We go on to show that they remove illumination artifacts using a mixture of synthetic and real-life images. This effect is highlighted more drastically when the illumination and exposure of the corresponding images is not the same.
Robust Calculation of Ego-Vehicle Corridors
(Computer Science Department, The University of Auckland, New Zealand, 2009) Jiang, Ruyi; Klette, Reinhard; Wang, Shigang; Vaudrey, Tobi
An important component of driver assistance systems (DAS) is lane detection, and has been studied since the 1990s. However, improving and generalizing lane detection solutions remains to be a challenging task until recently. A (physical) lane is defined by road boundaries or various kinds of lane marks, and this is only partially applicable for modeling the space an ego-vehicle is able to driving in. This paper proposes a concept of a (virtual) corridor for modeling this space. A corridor depends on information available about the motion of the ego-vehicle, as well as about the (physical) lane. This paper suggests robust corridor detection using hypothesis testing based on maximum a posterior (MAP) estimation. Then, boundary selection and road patch extension are applied as post-processing. Furthermore, a simple but efficient corridor tracking method is also discussed. This paper also informs the readers about experiments using images of some challenging road situations illustrating the usefulness of the proposed corridor detection and tracking scheme.
The Naked Truth about Cost Functions
(Computer Science Department, The University of Auckland, New Zealand, 2009) Hermann, Simon; Klette, Reinhard
This paper reports about the performance of various cost functions as common or possible for stereo matching, within a `purist's matching strategy', which does not integrate any further optimization approach such as a continuity term, coarse-to- ne, left/right consistency checks, or others. The performed experiments support a few general conclusions, such as about the relation between 1D- or 2D-window based cost functions, the use of zero-mean normalization, the census cost function, or the use of B-splines for subpixel accurate cost calculation.
A Study on Stereo and Motion Data Accuracy for a Moving Platform
(Computer Science Department, The University of Auckland, New Zealand, 2009) Morales, Sandino; Woo, Young Woo; Klette, Reinhard; Vaudrey, Tobi
Stereo and motion analysis are potential techniques for providing information for control or assistance systems in various robotics or driver assistance applications. This paper evaluates the performance of several stereo and motion algorithms over a long synthetic sequence (100 stereo pairs). Such an evaluation of low-level computer vision algorithms is necessary, as moving platforms are being used for image analysis in a wide area of applications. In this paper algorithms are evaluated with respect to robustness by modifying the test sequence with various types of realistic noise. The novelty of this paper is comparing top performing algorithms on a long sequence of images, taken from a moving platform.
An In-Depth Robustness Evaluation of Stereo Algorithms on Long Stereo Sequences
(Computer Science Department, The University of Auckland, New Zealand, 2008) Morales, Sandino; Vaudrey, Tobi; Klette, Reinhard
This paper presents an approach to test stereo algorithms against long stereo sequences (100+ image pairs). Stereo sequences of this length have not been quantitatively evaluated in the past. Using stereo sequences allows one to exploit the temporal information, which is in general not well used currently. Furthermore, the presented approach focuses on evaluating the robustness of algorithms against differing noise parameters (Gaussian noise, brightness differences, and blurring).
Affine Shape Adaptation of Blobs Moving in 3D Space
(Computer Science Department, The University of Auckland, New Zealand, 2008) Sanchez, Jorge A.; Destefanis, Eduardo; Klette, Reinhard; Morales, Sandino
The paper studies new constraints that characterize a 3D–motion field as observed from the relative motion of a camera. Such constraints are derived from the relative change in size of observed local image regions over time. To consider the image distortions that arise in a projective camera, a modified affine shape adaptation scheme is proposed for the case of blob detection, with an emphasis on robustness under important viewpoint changes and changes in lighting conditions. The resulting features and constraints are used to characterize the motion of an ego-vehicle by means of their navigation angles. We present results on synthetic as well as on real-world image sequences.
Dynamic Programming Stereo on Real-World Sequences
(Computer Science Department, The University of Auckland, New Zealand, 2008) Liu, Zhifeng; Klette, Reinhard
This paper proposes a way to approximate ground truth for real-world stereo sequences, and applies this for evaluating the performance of di erent variants of dynamic programming stereo analysis. This illustrates a way of performance evaluation, also allowing to derive sequence analysis diagrams. Obtained results di er from those obtained for the discussed algorithms on smaller, or engineered test data. This also shows the value of real-world testing.
Evaluation of Moving Object Segmentation Comparing 6D-Vision and Monocular Motion Constraints
(Computer Science Department, The University of Auckland, New Zealand, 2008) Vaudrey, Tobi; Wedel, Andreas; Rabe, Clemens; Klappstein, Jens; Klette, Reinhard
Detecting moving objects is a very important aspect of driver assistance systems (DAS). This paper handles this issue by using a vision based system mounted within the vehicle. The pipeline for both a stereoscopic and monocular approach are covered. Both approaches use image sequences and compare moving feature points over time. This sparse information is then segmented using the optimal graph-cut algorithm, by also incorporating the grey-scale images. This paper then evaluates and contrasts the two approaches to identify the accuracy and robustness of each approach. The two methods both work in real-time on normal PC hardware (Quad Core CPU).

Browse

Recent Submissions