Probabilistic visual recognition of artificial landmarks for simultaneous localization and mapping

Probabilistic robotics, most often applied to the problem of simultaneous localisation and mapping (SLAM), requires measures of uncertainly to accompany observations of the environment. This paper describes how uncertainly can be characterised for a vision system that locates coloured landmark in a typical laboratory environment. The paper describes a model of the uncertainly in segmentation, the internal camera model and the mounting of the camera on the robot. It =plains the implementation of the system on a laboratory robot, and provides experimental results that show the coherence of the uncertainly model,

Availability of advance visual information constrains association-football goalkeeping performance during penalty kicks

A pressing concern within the literature on anticipatory perceptual-motor behaviour is the lack of clarity on the applicability of data, observed under video-simulation task constraints, to actual performance in which actions are coupled to perception, as captured during in-situ experimental conditions. We developed an in-situ experimental paradigm which manipulated the duration of anticipatory visual information from a penalty taker’s actions to examine experienced goalkeepers’ vulnerability to deception for the penalty kick in association football. Irrespective of the penalty taker’s kick strategy, goalkeepers initiated movement responses earlier across consecutively earlier presentation points. Overall goalkeeping performance was better in non-deception trials than in deception conditions. In deception trials, the kinematic information presented up until the penalty taker initiated his/her kicking action had a negative effect on goalkeepers’ performance. It is concluded that goalkeepers are likely to benefit from not anticipating a penalty taker’s performance outcome based on information from the run-up, in preference to later information that emerges just before the initiation of the penalty taker’s kicking action.

The delta-phase spectrum with application to voice activity detection and speaker recognition

For several reasons, the Fourier phase domain is less favored than the magnitude domain in signal processing and modeling of speech. To correctly analyze the phase, several factors must be considered and compensated, including the effect of the step size, windowing function and other processing parameters. Building on a review of these factors, this paper investigates a spectral representation based on the Instantaneous Frequency Deviation, but in which the step size between processing frames is used in calculating phase changes, rather than the traditional single sample interval. Reflecting these longer intervals, the term delta-phase spectrum is used to distinguish this from instantaneous derivatives. Experiments show that mel-frequency cepstral coefficients features derived from the delta-phase spectrum (termed Mel-Frequency delta-phase features) can produce broadly similar performance to equivalent magnitude domain features for both voice activity detection and speaker recognition tasks. Further, it is shown that the fusion of the magnitude and phase representations yields performance benefits over either in isolation.

Robust suspicious behaviour detection for smart surveillance systems

Video surveillance technology, based on Closed Circuit Television (CCTV) cameras, is one of the fastest growing markets in the field of security technologies. However, the existing video surveillance systems are still not at a stage where they can be used for crime prevention. The systems rely heavily on human observers and are therefore limited by factors such as fatigue and monitoring capabilities over long periods of time. To overcome this limitation, it is necessary to have “intelligent” processes which are able to highlight the salient data and filter out normal conditions that do not pose a threat to security. In order to create such intelligent systems, an understanding of human behaviour, specifically, suspicious behaviour is required. One of the challenges in achieving this is that human behaviour can only be understood correctly in the context in which it appears. Although context has been exploited in the general computer vision domain, it has not been widely used in the automatic suspicious behaviour detection domain. So, it is essential that context has to be formulated, stored and used by the system in order to understand human behaviour. Finally, since surveillance systems could be modeled as largescale data stream systems, it is difficult to have a complete knowledge base. In this case, the systems need to not only continuously update their knowledge but also be able to retrieve the extracted information which is related to the given context. To address these issues, a context-based approach for detecting suspicious behaviour is proposed. In this approach, contextual information is exploited in order to make a better detection. The proposed approach utilises a data stream clustering algorithm in order to discover the behaviour classes and their frequency of occurrences from the incoming behaviour instances. Contextual information is then used in addition to the above information to detect suspicious behaviour. The proposed approach is able to detect observed, unobserved and contextual suspicious behaviour. Two case studies using video feeds taken from CAVIAR dataset and Z-block building, Queensland University of Technology are presented in order to test the proposed approach. From these experiments, it is shown that by using information about context, the proposed system is able to make a more accurate detection, especially those behaviours which are only suspicious in some contexts while being normal in the others. Moreover, this information give critical feedback to the system designers to refine the system. Finally, the proposed modified Clustream algorithm enables the system to both continuously update the system’s knowledge and to effectively retrieve the information learned in a given context. The outcomes from this research are: (a) A context-based framework for automatic detecting suspicious behaviour which can be used by an intelligent video surveillance in making decisions; (b) A modified Clustream data stream clustering algorithm which continuously updates the system knowledge and is able to retrieve contextually related information effectively; and (c) An update-describe approach which extends the capability of the existing human local motion features called interest points based features to the data stream environment.

Activity modelling in crowded environments : a soft decision approach

Probabilistic topic models have recently been used for activity analysis in video processing, due to their strong capacity to model both local activities and interactions in crowded scenes. In those applications, a video sequence is divided into a collection of uniform non-overlaping video clips, and the high dimensional continuous inputs are quantized into a bag of discrete visual words. The hard division of video clips, and hard assignment of visual words leads to problems when an activity is split over multiple clips, or the most appropriate visual word for quantization is unclear. In this paper, we propose a novel algorithm, which makes use of a soft histogram technique to compensate for the loss of information in the quantization process; and a soft cut technique in the temporal domain to overcome problems caused by separating an activity into two video clips. In the detection process, we also apply a soft decision strategy to detect unusual events.We show that the proposed soft decision approach outperforms its hard decision counterpart in both local and global activity modelling.

Improved sensorimotor adaptation after exhaustive exercise is accompanied by altered brain activity

Acute exercise has been shown to exhibit different effects on human sensorimotor behavior; however, the causes and mechanisms of the responses are often not clear. The primary aim of the present study was to determine the effects of incremental running until exhaustion on sensorimotor performance and adaptation in a tracking task. Subjects were randomly assigned to a running group (RG), a tracking group (TG), or a running followed by tracking group (RTG), with 10 subjects assigned to each group. Treadmill running velocity was initially set at 2.0 m s− 1, increasing by 0.5 m s− 1 every 5 min until exhaustion. Tracking consisted of 35 episodes (each 40 s) where the subjects' task was to track a visual target on a computer screen while the visual feedback was veridical (performance) or left-right reversed (adaptation). Resting electroencephalographic (EEG) activity was recorded before and after each experimental condition (running, tracking, rest). Tracking performance and the final amount of adaptation did not differ between groups. However, task adaptation was significantly faster in RTG compared to TG. In addition, increased alpha and beta power were observed following tracking in TG but not RTG although exhaustive running failed to induce significant changes in these frequency bands. Our results suggest that exhaustive running can facilitate adaptation processes in a manual tracking task. Attenuated cortical activation following tracking in the exercise condition was interpreted to indicate cortical efficiency and exercise-induced facilitation of selective central processes during actual task demands.

3D ellipsoid fitting for multi-view gait recognition

Gait recognition approaches continue to struggle with challenges including view-invariance, low-resolution data, robustness to unconstrained environments, and fluctuating gait patterns due to subjects carrying goods or wearing different clothes. Although computationally expensive, model based techniques offer promise over appearance based techniques for these challenges as they gather gait features and interpret gait dynamics in skeleton form. In this paper, we propose a fast 3D ellipsoidal-based gait recognition algorithm using a 3D voxel model derived from multi-view silhouette images. This approach directly solves the limitations of view dependency and self-occlusion in existing ellipse fitting model-based approaches. Voxel models are segmented into four components (left and right legs, above and below the knee), and ellipsoids are fitted to each region using eigenvalue decomposition. Features derived from the ellipsoid parameters are modeled using a Fourier representation to retain the temporal dynamic pattern for classification. We demonstrate the proposed approach using the CMU MoBo database and show that an improvement of 15-20% can be achieved over a 2D ellipse fitting baseline.

Automated recognition of drunk driving on highways from video sequences

A new method for the detection of abnormal vehicle trajectories is proposed. It couples optical flow extraction of vehicle velocities with a neural network classifier. Abnormal trajectories are indicative of drunk or sleepy drivers. A single feature of the vehicle, eg., a tail light, is isolated and the optical flow computed only around this feature rather than at each pixel in the image.

Feature-based visual odometry and featureless place recognition for SLAM in 2.5D environments

In this paper we present a novel algorithm for localization during navigation that performs matching over local image sequences. Instead of calculating the single location most likely to correspond to a current visual scene, the approach finds candidate matching locations within every section (subroute) of all learned routes. Through this approach, we reduce the demands upon the image processing front-end, requiring it to only be able to correctly pick the best matching image from within a short local image sequence, rather than globally. We applied this algorithm to a challenging downhill mountain biking visual dataset where there was significant perceptual or environment change between repeated traverses of the environment, and compared performance to applying the feature-based algorithm FAB-MAP. The results demonstrate the potential for localization using visual sequences, even when there are no visual features that can be reliably detected.

Face recognition across pose on video using eigen light-fields

We propose an approach to employ eigen light-fields for face recognition across pose on video. Faces of a subject are collected from video frames and combined based on the pose to obtain a set of probe light-fields. These probe data are then projected to the principal subspace of the eigen light-fields within which the classification takes place. We modify the original light-field projection and found that it is more robust in the proposed system. Evaluation on VidTIMIT dataset has demonstrated that the eigen light-fields method is able to take advantage of multiple observations contained in the video.

Engaging bodily with video in design

Video is commonly used as a method for recording embodied interaction for purposes of analysis and design and has been proposed as a useful ‘material’ for interaction designers to engage with. But video is not a straight forward reproduction of embodied activity – in themselves video recordings ‘flatten’ the space of embodied interaction, they impose a perspective on unfolding action, and remove the embodied spatial and social context within which embodied interaction unfolds. This does not mean that video is not a useful medium with which to engage as part of a process of investigating and designing for embodied interaction – but crucially, it requires that as people attempting to engage with video, designers own bodies and bodily understandings must be engaged with and brought into play. This paper describes and reflects upon our experiences of engaging with video in two different activities as part of a larger research project investigating the design of gestural interfaces for a dental surgery context.

Probabilistic matching of image sets for video-based face recognition

We address the problem of face recognition on video by employing the recently proposed probabilistic linear discrimi-nant analysis (PLDA). The PLDA has been shown to be robust against pose and expression in image-based face recognition. In this research, the method is extended and applied to video where image set to image set matching is performed. We investigate two approaches of computing similarities between image sets using the PLDA: the closest pair approach and the holistic sets approach. To better model face appearances in video, we also propose the heteroscedastic version of the PLDA which learns the within-class covariance of each individual separately. Our experi-ments on the VidTIMIT and Honda datasets show that the combination of the heteroscedastic PLDA and the closest pair approach achieves the best performance.