Warping trajectories for video synchronization

Temporal synchronization of multiple video recordings of the same dynamic event is a critical task in many computer vision applications e.g. novel view synthesis and 3D reconstruction. Typically this information is implied, since recordings are made using the same timebase, or time-stamp information is embedded in the video streams. Recordings using consumer grade equipment do not contain this information; hence, there is a need to temporally synchronize signals using the visual information itself. Previous work in this area has either assumed good quality data with relatively simple dynamic content or the availability of precise camera geometry. In this paper, we propose a technique which exploits feature trajectories across views in a novel way, and specifically targets the kind of complex content found in consumer generated sports recordings, without assuming precise knowledge of fundamental matrices or homographies. Our method automatically selects the moving feature points in the two unsynchronized videos whose 2D trajectories can be best related, thereby helping to infer the synchronization index. We evaluate performance using a number of real recordings and show that synchronization can be achieved to within 1 sec, which is better than previous approaches. Copyright 2013 ACM.

An experimental study on the load carrying capacity of cold-formed steel studs and panels

A full-scale experimental study on the structural performance of load-bearing wall panels made of cold-formed steel frames and boards is presented. Six different types of C-channel stud, a total of 20 panels with one middle stud and 10 panels with two middle studs were tested under vertical compression until failure. For panels, the main variables considered are screw spacing (300 mm, 400 mm, or 600 mm) in the middle stud, board type (oriented strand board - OSB, cement particle board - CPB, or calcium silicate board - CSB), board number (no sheathing, one-side sheathing, or two-side sheathing), and loading type (1, 3, or 4-point loading). The measured load capacity of studs and panels agrees well with analytical prediction. Due to the restraint by rivet connections between stud and track, the effective length factor for the middle stud and the side stud in a frame (unsheathed panel) is reduced to 0.90 and 0.84, respectively. The load carrying capacity of a stud increases significantly whenever one- or two-side sheathing is used, although the latter is significantly more effective. It is also dependent upon the type of board used. Whereas panels with either OSB or CPB boards have nearly identical load carrying capacity, panels with CSB boards are considerably weaker. Screw spacing affects the load carrying capacity of a stud. When the screw spacing on the middle stud in panels with one-side sheathing is reduced from 600 mm to 300 mm, its load carrying capacity increases by 14.5 %, 20.6% and 94.2% for OSB, CPB and CSB, respectively.

SPEECH RECOGNITION IN VODIS II.

VODIS II, a research system in which recognition is based on the conventional one-pass connected-word algorithm extended in two ways, is described. Syntactic constraints can now be applied directly via context-free-grammar rules, and the algorithm generates a lattice of candidate word matches rather than a single globally optimal sequence. This lattice is then processed by a chart parser and an intelligent dialogue controller to obtain the most plausible interpretations of the input. A key feature of the VODIS II architecture is that the concept of an abstract word model allows the system to be used with different pattern-matching technologies and hardware. The current system implements the word models on a real-time dynamic-time-warping recognizer.

Warped Mixtures for Nonparametric Cluster Shapes

A mixture of Gaussians fit to a single curved or heavy-tailed cluster will report that the data contains many clusters. To produce more appropriate clusterings, we introduce a model which warps a latent mixture of Gaussians to produce nonparametric cluster shapes. The possibly low-dimensional latent mixture model allows us to summarize the properties of the high-dimensional clusters (or density manifolds) describing the data. The number of manifolds, as well as the shape and dimension of each manifold is automatically inferred. We derive a simple inference scheme for this model which analytically integrates out both the mixture parameters and the warping function. We show that our model is effective for density estimation, performs better than infinite Gaussian mixture models at recovering the true number of clusters, and produces interpretable summaries of high-dimensional datasets.

Effects of solution exposure on the combined axial-shear behaviour of unidirectional CFRP rods

In fibre reinforced polymer (FRP) prestressed concrete applications, an FRP tendon must sustain high axial tensile stresses and, if cracks occur, additional dowel forces. The tendon may also be exposed to solutions and so the combined axial-shear stress performance after long-term environmental exposure is important. Experiments were conducted to investigate the combined axial-shear stress failure envelope for unidirectional carbon FRP tendons which had been exposed to either water, salt water or concrete pore solution at 60 °C for approximately 18 months. The underlying load resisting mechanisms were found to depend on the loading configuration, restraint effects and the initial stress state. When saturated, CFRP tendons are likely to exhibit a reduced shear stiffness. However, the ultimate limit state appeared to be fibre-dominated and was therefore less susceptible to reductions due to solution uptake effects. © 2012 Elsevier Ltd. All rights reserved.

Learning based automatic face annotation for arbitrary poses and expressions from frontal images only

Statistical approaches for building non-rigid deformable models, such as the Active Appearance Model (AAM), have enjoyed great popularity in recent years, but typically require tedious manual annotation of training images. In this paper, a learning based approach for the automatic annotation of visually deformable objects from a single annotated frontal image is presented and demonstrated on the example of automatically annotating face images that can be used for building AAMs for fitting and tracking. This approach employs the idea of initially learning the correspondences between landmarks in a frontal image and a set of training images with a face in arbitrary poses. Using this learner, virtual images of unseen faces at any arbitrary pose for which the learner was trained can be reconstructed by predicting the new landmark locations and warping the texture from the frontal image. View-based AAMs are then built from the virtual images and used for automatically annotating unseen images, including images of different facial expressions, at any random pose within the maximum range spanned by the virtually reconstructed images. The approach is experimentally validated by automatically annotating face images from three different databases. © 2009 IEEE.