Along with technological developments and increasing competition there is a major incentive for companies to produce and market high quality products before their competitors. In order to conquer a bigger portion of the market share, companies have to ensure the quality of the product in a shorter time frame. To accomplish this task companies try to automate their test processes as much as possible. It is critical to investigate and understand the problems that occur during different stages of test automation processes. In this paper we report on a case study on automatic analysis of non-functional test results. We discuss challenges in the face of continuous integration and deployment and provide improvement suggestions based on interviews at a large company in Sweden. The key contributions of this work are filling the knowledge gap in research about performance regression test analysis automation and providing warning signs and a road map for the industry.

We propose a novel method for iterative learning of point correspondences between image sequences. Points moving on surfaces in 3D space are projected into two images. Given a point in either view, the considered problem is to determine the corresponding location in the other view. The geometry and distortions of the projections are unknown, as is the shape of the surface. Given several pairs of point sets but no access to the 3D scene, correspondence mappings can be found by excessive global optimization or by the fundamental matrix if a perspective projective model is assumed. However, an iterative solution on sequences of point-set pairs with general imaging geometry is preferable. We derive such a method that optimizes the mapping based on Neyman's chi-square divergence between the densities representing the uncertainties of the estimated and the actual locations. The densities are represented as channel vectors computed with a basis function approach. The mapping between these vectors is updated with each new pair of images such that fast convergence and high accuracy are achieved. The resulting algorithm runs in real time and is superior to state-of-the-art methods in terms of convergence and accuracy in a number of experiments.

A common computer vision task is navigation and mapping. Many indoor navigation tasks require depth knowledge of flat, unstructured surfaces (walls, floor, ceiling). With passive illumination only, this is an ill-posed problem. Inspired by small children using a torchlight, we use a spotlight for active illumination. Using our torchlight approach, depth and orientation estimation of unstructured, flat surfaces boils down to estimation of ellipse parameters. The extraction of ellipses is very robust and requires little computational effort.

During the past few years, a number of verification tools have been developed for real-time systems in the framework of timed automata (e.g. KRONOS and UPPAAL). One of the major problems in applying these tools to industrial-size systems is the huge memory-usage for the exploration of the state-space of a network (or product) of timed automata, as the model-checkers must keep information on not only the control structure of the automata but also the clock values specified by clock constraints. In this paper, we present a compact data structure for representing clock constraints. The data structure is based on an O(n/sup 3/) algorithm which, given a constraint system over real-valued variables consisting of bounds on differences, constructs an equivalent system with a minimal number of constraints. In addition, we have developed an on-the-fly, reduction technique to minimize the space-usage. Based on static analysis of the control structure of a network of timed automata, we are able to compute a set of symbolic states that cover all the dynamic loops of the network in an on-the-fly searching algorithm, and thus ensure termination in reachability analysis. The two techniques and their combination have been implemented in the tool UPPAAL. Our experimental results demonstrate that the techniques result in truly significant space-reductions: for six examples from the literature, the space saving is between 75% and 94%, and in (nearly) all examples time-performance is improved. Also noteworthy is the observation that the two techniques are completely orthogonal.