An Event Driven Fusion Approach for Enjoyment Recognition in Real-time

Social signals and interpretation of carried information is of high importance in Human Computer Interaction. Often used for affect recognition, the cues within these signals are displayed in various modalities. Fusion of multi-modal signals is a natural and interesting way to improve automatic classification of emotions transported in social signals. Throughout most present studies, uni-modal affect recognition as well as multi-modal fusion, decisions are forced for fixed annotation segments across all modalities. In this paper, we investigate the less prevalent approach of event driven fusion, which indirectly accumulates asynchronous events in all modalities for final predictions. We present a fusion approach, handling short-timed events in a vector space, which is of special interest for real-time applications. We compare results of segmentation based uni-modal classification and fusion schemes to the event driven fusion approach. The evaluation is carried out via detection of enjoyment-episodes within the audiovisual Belfast Story-Telling Corpus.

Building pH Sensors into Paper-based Small-molecular Logic Sys-tems for Very Simple Detection of Edges of Objects

Genetically-engineered bacteria and reactive DNA networks detect edges of objects, as done in our retinas and as also found within computer vision. We now demonstrate that simple molecular logic systems (a combination of a pH sensor, a photo acid generator and a pH buffer spread on paper) without any organization can achieve this relatively complex computational goal with good-fidelity. This causes a jump in the complexity achievable by molecular logic-based computation and extends its applicability. The molecular species involved in light dose-driven 'off-on-off' fluorescence is diverted in the ‘on’ state by proton diffusion from irradiated to unirradiated regions where it escapes a strong quencher, thus visualizing the edge of a mask.

Evidence reasoning for event inference in smart transport video surveillance

In this paper we present a new event recognition framework, based on the Dempster-Shafer theory of evidence, which combines the evidence from multiple atomic events detected by low-level computer vision analytics. The proposed framework employs evidential network modelling of composite events. This approach can effectively handle the uncertainty of the detected events, whilst inferring high-level events that have semantic meaning with high degrees of belief. Our scheme has been comprehensively evaluated against various scenarios that simulate passenger behaviour on public transport platforms such as buses and trains. The average accuracy rate of our method is 81% in comparison to 76% by a standard rule-based method.

Enhancing Linear Programming with Motion Modeling for Multi-target Tracking

In this paper we extend the minimum-cost network flow approach to multi-target tracking, by incorporating a motion model, allowing the tracker to better cope with longterm occlusions and missed detections. In our new method, the tracking problem is solved iteratively: Firstly, an initial tracking solution is found without the help of motion information. Given this initial set of tracklets, the motion at each detection is estimated, and used to refine the tracking solution.
Finally, special edges are added to the tracking graph, allowing a further revised tracking solution to be found, where distant tracklets may be linked based on motion similarity. Our system has been tested on the PETS S2.L1 and Oxford town-center sequences, outperforming the baseline system, and achieving results comparable with the current state of the art.

Video Event Recognition by Dempster-Shafer Theory

This paper presents an event recognition framework, based on Dempster-Shafer theory, that combines evidence of events from low-level computer vision analytics. The proposed method employing evidential network modelling of composite events, is able to represent uncertainty of event output from low level video analysis and infer high level events with semantic meaning along with degrees of belief. The method has been evaluated on videos taken of subjects entering and leaving a seated area. This has relevance to a number of transport scenarios, such as onboard buses and trains, and also in train stations and airports. Recognition results of 78% and 100% for four composite events are encouraging.

Region-based Mixture of Gaussians Modelling for Foreground Detection in Dynamic Scenes

One of the most widely used techniques in computer vision for foreground detection is to model each background pixel as a Mixture of Gaussians (MoG). While this is effective for a static camera with a fixed or a slowly varying background, it fails to handle any fast, dynamic movement in the background. In this paper, we propose a generalised framework, called region-based MoG (RMoG), that takes into consideration neighbouring pixels while generating the model of the observed scene. The model equations are derived from Expectation Maximisation theory for batch mode, and stochastic approximation is used for online mode updates. We evaluate our region-based approach against ten sequences containing dynamic backgrounds, and show that the region-based approach provides a performance improvement over the traditional single pixel MoG. For feature and region sizes that are equal, the effect of increasing the learning rate is to reduce both true and false positives. Comparison with four state-of-the art approaches shows that RMoG outperforms the others in reducing false positives whilst still maintaining reasonable foreground definition. Lastly, using the ChangeDetection (CDNet 2014) benchmark, we evaluated RMoG against numerous surveillance scenes and found it to amongst the leading performers for dynamic background scenes, whilst providing comparable performance for other commonly occurring surveillance scenes.

New FMO Type to Flag ROI in H264/AVC

This paper presents a new type of Flexible Macroblock Ordering (FMO) type for the H.264 Advanced Video Coding (AVC) standard, which can more efficiently flag the position and shape of regions of interest (ROIs) in each frame. In H.264/AVC, 7 types of FMO have been defined, all of which are designed for error resilience. Most previous work related to ROI processing has adopted Type-2 (foreground & background), or Type-6 (explicit), to flag the position and shape of the ROI. However, only rectangular shapes are allowed in Type-2 and for non-rectangular shapes, the non-ROI macroblocks may be wrongly flagged as being within the ROI, which could seriously affect subsequent processing of the ROI. In Type-6, each macroblock in a frame uses fixed-length bits to indicate to its slice group. In general, each ROI is assigned to one slice group identity. Although this FMO type can more accurately flag the position and shape of the ROI, it incurs a significant bitrate overhead. The proposed new FMO type uses the smallest rectangle that covers the ROI to indicate its position and a spiral binary mask is employed within the rectangle to indicate the shape of the ROI. This technique can accurately flag the ROI and provide significantly savings in the bitrate overhead. Compared with Type-6, an 80% to 90% reduction in the bitrate overhead can be obtained while achieving the same accuracy.

Virtual Thrower Versus Real Goalkeeper: The Influence of Different Visual Conditions on Performance

In order to use virtual reality as a sport analysis tool, we need to be sure that an immersed athlete reacts realistically in a virtual environment. This has been validated for a real handball goalkeeper facing a virtual thrower. However, we currently ignore which visual variables induce a realistic motor behavior of the immersed handball goalkeeper. In this study, we used virtual reality to dissociate the visual information related to the movements of the player from the visual information related to the trajectory of the ball. Thus, the aim is to evaluate the relative influence of these different visual information sources on the goalkeeper's motor behavior. We tested 10 handball goalkeepers who had to predict the final position of the virtual ball in the goal when facing the following: only the throwing action of the attacking player (TA condition), only the resulting ball trajectory (BA condition), and both the throwing action of the attacking player and the resulting ball trajectory (TB condition). Here we show that performance was better in the BA and TB conditions, but contrary to expectations, performance was substantially worse in the TA condition. A significant effect of ball landing zone does, however, suggest that the relative importance between visual information from the player and the ball depends on the targeted zone in the goal. In some cases, body-based cues embedded in the throwing actions may have a minor influence on the ball trajectory and vice versa. Kinematics analysis was then combined with these results to determine why such differences occur depending on the ball landing zone and consequently how it can clarify the role of different sources of visual information on the motor behavior of an athlete immersed in a virtual environment.

Evidential event inference in transport video surveillance

This paper presents a new framework for multi-subject event inference in surveillance video, where measurements produced by low-level vision analytics usually are noisy, incomplete or incorrect. Our goal is to infer the composite events undertaken by each subject from noise observations. To achieve this, we consider the temporal characteristics of event relations and propose a method to correctly associate the detected events with individual subjects. The Dempster–Shafer (DS) theory of belief functions is used to infer events of interest from the results of our vision analytics and to measure conflicts occurring during the event association. Our system is evaluated against a number of videos that present passenger behaviours on a public transport platform namely buses at different levels of complexity. The experimental results demonstrate that by reasoning with spatio-temporal correlations, the proposed method achieves a satisfying performance when associating atomic events and recognising composite events involving multiple subjects in dynamic environments.

PatchCity: Procedural City Generation using Texture Synthesis

PatchCity is a new approach to the procedural generation of city models. The algorithm uses texture synthesis to create a city layout in the visual style of one or more input examples. Data is provided in vector graphic form from either real or synthetic city definitions. The paper describes the PatchCity algorithm, illustrates its use, and identifies its strengths and limitations. The technique provides a greater range of features and styles of city layout than existing generative methods, thereby achieving results that are more realistic. An open source implementation of the algorithm is available.

Investigation into DCT Feature Selection for Visual Lip-Based Biometric Authentication

This paper investigated using lip movements as a behavioural biometric for person authentication. The system was trained, evaluated and tested using the XM2VTS dataset, following the Lausanne Protocol configuration II. Features were selected from the DCT coefficients of the greyscale lip image. This paper investigated the number of DCT coefficients selected, the selection process, and static and dynamic feature combinations. Using a Gaussian Mixture Model - Universal Background Model framework an Equal Error Rate of 2.20% was achieved during evaluation and on an unseen test set a False Acceptance Rate of 1.7% and False Rejection Rate of 3.0% was achieved. This compares favourably with face authentication results on the same dataset whilst not being susceptible to spoofing attacks.

Kernelized structural classification for 3D dogs body parts detection

Despite pattern recognition methods for human behavioral analysis has flourished in the last decade, animal behavioral analysis has been almost neglected. Those few approaches are mostly focused on preserving livestock economic value while attention on the welfare of companion animals, like dogs, is now emerging as a social need. In this work, following the analogy with human behavior recognition, we propose a system for recognizing body parts of dogs kept in pens. We decide to adopt both 2D and 3D features in order to obtain a rich description of the dog model. Images are acquired using the Microsoft Kinect to capture the depth map images of the dog. Upon depth maps a Structural Support Vector Machine (SSVM) is employed to identify the body parts using both 3D features and 2D images. The proposal relies on a kernelized discriminative structural classificator specifically tailored for dogs independently from the size and breed. The classification is performed in an online fashion using the LaRank optimization technique to obtaining real time performances. Promising results have emerged during the experimental evaluation carried out at a dog shelter, managed by IZSAM, in Teramo, Italy.

Kinect and Episodic Reasoning for Human Action Recognition

This paper presents a method for rational behaviour recognition that combines vision-based pose estimation with knowledge modeling and reasoning. The proposed method consists of two stages. First, RGB-D images are used in the estimation of the body postures. Then, estimated actions are evaluated to verify that they make sense. This method requires rational behaviour to be exhibited. To comply with this requirement, this work proposes a rational RGB-D dataset with two types of sequences, some for training and some for testing. Preliminary results show the addition of knowledge modeling and reasoning leads to a significant increase of recognition accuracy when compared to a system based only on computer vision.

Recurrent Convolutional Network for Video-based Person Re-Identification

In this paper we propose a novel recurrent neural networkarchitecture for video-based person re-identification.Given the video sequence of a person, features are extracted from each frame using a convolutional neural network that incorporates a recurrent final layer, which allows information to flow between time-steps. The features from all time steps are then combined using temporal pooling to give an overall appearance feature for the complete sequence. The convolutional network, recurrent layer, and temporal pooling layer, are jointly trained to act as a feature extractor for video-based re-identification using a Siamese network architecture.Our approach makes use of colour and optical flow information in order to capture appearance and motion information which is useful for video re-identification. Experiments are conduced on the iLIDS-VID and PRID-2011 datasets to show that this approach outperforms existing methods of video-based re-identification.

https://github.com/niallmcl/Recurrent-Convolutional-Video-ReID
Project Source Code

Learning from Outdoor Webcams: Surveillance of Physical Activity Across Environments

Abstract
Publicly available, outdoor webcams continuously view the world and share images. These cameras include traffic cams, campus cams, ski-resort cams, etc. The Archive of Many Outdoor Scenes (AMOS) is a project aiming to geolocate, annotate, archive, and visualize these cameras and images to serve as a resource for a wide variety of scientific applications. The AMOS dataset has archived over 750 million images of outdoor environments from 27,000 webcams since 2006. Our goal is to utilize the AMOS image dataset and crowdsourcing to develop reliable and valid tools to improve physical activity assessment via online, outdoor webcam capture of global physical activity patterns and urban built environment characteristics.
This project’s grand scale-up of capturing physical activity patterns and built environments is a methodological step forward in advancing a real-time, non-labor intensive assessment using webcams, crowdsourcing, and eventually machine learning. The combined use of webcams capturing outdoor scenes every 30 min and crowdsources providing the labor of annotating the scenes allows for accelerated public health surveillance related to physical activity across numerous built environments. The ultimate goal of this public health and computer vision collaboration is to develop machine learning algorithms that will automatically identify and calculate physical activity patterns.