Real-time video event detection in crowded scenes using MPEG derived features : a multiple instance learning approach

This paper presents an investigation into event detection in crowded scenes, where the event of interest co-occurs with other activities and only binary labels at the clip level are available. The proposed approach incorporates a fast feature descriptor from the MPEG domain, and a novel multiple instance learning (MIL) algorithm using sparse approximation and random sensing. MPEG motion vectors are used to build particle trajectories that represent the motion of objects in uniform video clips, and the MPEG DCT coefficients are used to compute a foreground map to remove background particles. Trajectories are transformed into the Fourier domain, and the Fourier representations are quantized into visual words using the K-Means algorithm. The proposed MIL algorithm models the scene as a linear combination of independent events, where each event is a distribution of visual words. Experimental results show that the proposed approaches achieve promising results for event detection compared to the state-of-the-art.

Nonrigid face alignment for unknown subject in video

Non-rigid face alignment is a very important task in a large range of applications but the existing tracking based non-rigid face alignment methods are either inaccurate or requiring person-specific model. This dissertation has developed simultaneous alignment algorithms that overcome these constraints and provide alignment with high accuracy, efficiency, robustness to varying image condition, and requirement of only generic model.

Clinical skin examination outcomes after a video-based behavioral intervention analysis from a randomized clinical trial

Importance Older men are at risk of dying of melanoma. Objective To assess attendance at and clinical outcomes of clinical skin examinations (CSEs) in older men exposed to a video-based behavioral intervention. Design, Setting, and Participants This was a behavioral randomized clinical trial of a video-based intervention in men aged at least 50 years. Between June 1 and August 31, 2008, men were recruited, completed baseline telephone interviews, and were than randomized to receive either a video-based intervention (n = 469) or brochures only (n = 461; overall response rate, 37.1%) and were again interviewed 7 months later (n = 870; 93.5% retention). Interventions Video on skin self-examination and skin awareness and written informational materials. The control group received written materials only. Main Outcomes and Measures Participants who reported a CSE were asked for the type of CSE (skin spot, partial body, or whole body), who initiated it, whether the physician noted any suspicious lesions, and, if so, how lesions were managed. Physicians completed a case report form that included the type of CSE, who initiated it, the number of suspicious lesions detected, how lesions were managed (excision, nonsurgical treatment, monitoring, or referral), and pathology reports after lesion excision or biopsy. Results Overall, 540 of 870 men (62.1%) self-reported a CSE since receiving intervention materials, and 321 of 540 (59.4%) consented for their physician to provide medical information (received for 266 of 321 [82.9%]). Attendance of any CSE was similar between groups (intervention group, 246 of 436 [56.4%]; control group, 229 of 434 [52.8%]), but men in the intervention group were more likely to self-report a whole-body CSE (154 of 436 [35.3%] vs 118 of 434 [27.2%] for control group; P = .01). Two melanomas, 29 squamous cell carcinomas, and 38 basal cell carcinomas were diagnosed, with a higher proportion of malignant lesions in the intervention group (60.0% vs 40.0% for controls; P = .03). Baseline attitudes, behaviors, and skin cancer history were associated with higher odds of CSE and skin cancer diagnosis. Conclusions and Relevance A video-based intervention may increase whole-body CSE and skin cancer diagnosis in older men. Trial Registration: anzctr.org.au Identifier: ACTRN12608000384358

Quality based frame selection for face clustering in news video

Clustering identities in a broadcast video is a useful task to aid in video annotation and retrieval. Quality based frame selection is a crucial task in video face clustering, to both improve the clustering performance and reduce the computational cost. We present a frame work that selects the highest quality frames available in a video to cluster the face. This frame selection technique is based on low level and high level features (face symmetry, sharpness, contrast and brightness) to select the highest quality facial images available in a face sequence for clustering. We also consider the temporal distribution of the faces to ensure that selected faces are taken at times distributed throughout the sequence. Normalized feature scores are fused and frames with high quality scores are used in a Local Gabor Binary Pattern Histogram Sequence based face clustering system. We present a news video database to evaluate the clustering system performance. Experiments on the newly created news database show that the proposed method selects the best quality face images in the video sequence, resulting in improved clustering performance.

Image segmentation and pose estimation of humans in video

This thesis introduces improved techniques towards automatically estimating the pose of humans from video. It examines a complete workflow to estimating pose, from the segmentation of the raw video stream to extract silhouettes, to using the silhouettes in order to determine the relative orientation of parts of the human body. The proposed segmentation algorithms have improved performance and reduced complexity, while the pose estimation shows superior accuracy during difficult cases of self occlusion.

Patron-driven acquisition of online video

Letting the patron choose ebooks has been a successful experience. Why not apply the same purchase model to other formats? This showcase outlines Queensland University of Technology’s experience with a trial of patron driven acquisition (PDA) for online video. The trial commencing in August 2012 provided access to over 700 online videos licensed from Kanopy across a number of discipline areas. As online video publishing is still in the early stages of development, and as the trial is only in the very early stages, it is too early to draw any firm conclusions about the likely suitability of this model for online video selection and acquisition. However, the trial provides some interesting initial comparisons with ebook PDA and existing online video purchase models and prompts further consideration of PDA as a method for online video selection and licensing.

Possible, inevitable or fait accompli? An analysis of streaming video acquisition, acceptance and use in higher education

This paper explores the potential for online video as a mechanism to transform the ways students learn, as measured by research, user experience and usage following surveys and trials of patron-driven acquisition collaboratively undertaken by Queensland University of Technology, La Trobe University and Kanopy.

‘Teachers flipping out’ beyond the online lecture : maximising the educational potential of video

While video is recognised as an important medium for teaching and learning in the digital age, many video resources are not as effective as they might be, because they do not adequately exploit the strengths of the medium. Presented here are some case studies of video learning resources produced for various courses in a university environment. This ongoing project attempts to identify pedagogic strategies for the use of video; learning situations in which video has the most efficacy; and what production techniques can be employed to make effective video learning resources.

Semi-binary based video features for activity representation

Efficient and effective feature detection and representation is an important consideration when processing videos, and a large number of applications such as motion analysis, 3D scene understanding, tracking etc. depend on this. Amongst several feature description methods, local features are becoming increasingly popular for representing videos because of their simplicity and efficiency. While they achieve state-of-the-art performance with low computational complexity, their performance is still too limited for real world applications. Furthermore, rapid increases in the uptake of mobile devices has increased the demand for algorithms that can run with reduced memory and computational requirements. In this paper we propose a semi binary based feature detectordescriptor based on the BRISK detector, which can detect and represent videos with significantly reduced computational requirements, while achieving comparable performance to the state of the art spatio-temporal feature descriptors. First, the BRISK feature detector is applied on a frame by frame basis to detect interest points, then the detected key points are compared against consecutive frames for significant motion. Key points with significant motion are encoded with the BRISK descriptor in the spatial domain and Motion Boundary Histogram in the temporal domain. This descriptor is not only lightweight but also has lower memory requirements because of the binary nature of the BRISK descriptor, allowing the possibility of applications using hand held devices.We evaluate the combination of detectordescriptor performance in the context of action classification with a standard, popular bag-of-features with SVM framework. Experiments are carried out on two popular datasets with varying complexity and we demonstrate comparable performance with other descriptors with reduced computational complexity.

Acceptability-based QoE models for mobile video

Quality of experience (QoE) measures the overall perceived quality of mobile video delivery from subjective user experience and objective system performance. Current QoE computing models have two main limitations: 1) insufficient consideration of the factors influencing QoE, and; 2) limited studies on QoE models for acceptability prediction. In this paper, a set of novel acceptability-based QoE models, denoted as A-QoE, is proposed based on the results of comprehensive user studies on subjective quality acceptance assessments. The models are able to predict users’ acceptability and pleasantness in various mobile video usage scenarios. Statistical regression analysis has been used to build the models with a group of influencing factors as independent predictors, including encoding parameters and bitrate, video content characteristics, and mobile device display resolution. The performance of the proposed A-QoE models has been compared with three well-known objective Video Quality Assessment metrics: PSNR, SSIM and VQM. The proposed A-QoE models have high prediction accuracy and usage flexibility. Future user-centred mobile video delivery systems can benefit from applying the proposed QoE-based management to optimize video coding and quality delivery decisions.

Video analytics for the detection of near-miss incidents on approach to railway level crossings

Recent modelling of socio-economic costs by the Australian railway industry in 2010 has estimated the cost of level crossing accidents to exceed AU$116 million annually. To better understand causal factors that contribute to these accidents, the Cooperative Research Centre for Rail Innovation is running a project entitled Baseline Level Crossing Video. The project aims to improve the recording of level crossing safety data by developing an intelligent system capable of detecting near-miss incidents and capturing quantitative data around these incidents. To detect near-miss events at railway level crossings a video analytics module is being developed to analyse video footage obtained from forward-facing cameras installed on trains. This paper presents a vision base approach for the detection of these near-miss events. The video analytics module is comprised of object detectors and a rail detection algorithm, allowing the distance between a detected object and the rail to be determined. An existing publicly available Histograms of Oriented Gradients (HOG) based object detector algorithm is used to detect various types of vehicles in each video frame. As vehicles are usually seen from a sideway view from the cabin’s perspective, the results of the vehicle detector are verified using an algorithm that can detect the wheels of each detected vehicle. Rail detection is facilitated using a projective transformation of the video, such that the forward-facing view becomes a bird’s eye view. Line Segment Detector is employed as the feature extractor and a sliding window approach is developed to track a pair of rails. Localisation of the vehicles is done by projecting the results of the vehicle and rail detectors on the ground plane allowing the distance between the vehicle and rail to be calculated. The resultant vehicle positions and distance are logged to a database for further analysis. We present preliminary results regarding the performance of a prototype video analytics module on a data set of videos containing more than 30 different railway level crossings. The video data is captured from a journey of a train that has passed through these level crossings.

Talk about a YouTube video in preschool : the mutual production of shared understanding for learning with digital technology

Much of what is written about digital technologies in preschool contexts focuses on young children’s acquisition of skills rather than their meaning-making during use of technologies. In this paper, we consider how the viewing of a YouTube video was used by a teacher and children to produce shared understandings about it. Conversation analysis of talk and interaction during the viewing of the video establishes some of the ways that individual accounts of events were produced for others and then endorsed as shared understandings. The analysis establishes how adults and children made use of verbal and embodied actions during interactions to produce shared understandings of the YouTube video, the events it recorded and written commentary about those events

New directions for digital video creation in the classroom : spatiality, embodiment, and creativity

This column features a conversation (via email, image sharing, and Facetime) that took place over several months between two international theorists of digital filmmaking from schools in two countries—Professors Jason Ranker (Portland State University, Oregon, United States) and Kathy Mills (Queensland University of Technology, Australia). The authors discuss emerging ways of thinking about video making, sharing tips and anecdotes from classroom experience to inspire teachers to explore with adolescents the meaning potentials of digital video creation. The authors briefly discuss their previous work in this area, and then move into a discussion of how the material spaces in which students create videos profoundly shape the films' meanings and significance. The article ends with a discussion of how students can take up creative new directions, pushing the boundaries of the potentials of classroom video making and uncovering profound uses of the medium.

E-motion diaries : blogging emotional responses to online learning in higher education

Emotions are inherently social, and are central to learning, online interaction and literacy practices (Shen, Wang, & Shen, 2009). Demonstrating the dynamic sociality of literacy practice, we used e-motion diaries or web logs to explore the emotional states of pre-service high school teachers’ experiences of online learning activities. This is because the methods of communication used by university educators in online learning and writing environments play an important role in fulfilling students’ need for social interaction and inclusion (McInnerney & Roberts, 2004). Feelings of isolation and frustration are common emotions experienced by students in many online learning environments, and are associated with the success or failure of online interactions and learning (Su, et al., 2005). The purpose of the study was to answer the research question: What are the trajectories of pre-service teachers’ emotional states during online learning experiences? This is important because emotions are central to learning, and the current trend toward Massive Open Online Courses (MOOCs) needs research about students’ emotional connections in online learning environments (Kop, 2011). The project was conducted with a graduate class of 64 high school science pre-service teachers in Science Education Curriculum Studies in a large Australian university, including males and females from a variety of cultural backgrounds, aged 22-55 years. Online activities involved the students watching a series of streamed live lectures for the first 5 weeks providing a varied set of learning experiences, such as viewing science demonstrations (e.g., modeling the use of discrepant events). Each week, students provided feedback on learning by writing and posting an e-motion diary or web log about their emotional response. Students answered the question: What emotions did you experience during this learning experience? The descriptive data set included 284 online posts, with students contributing multiple entries. Linguistic appraisal theory, following Martin and White (2005), was used to regroup the 22 different discrete emotions reported by students into the six main affect groups – three positive and three negative: unhappiness/happiness, insecurity/security, and dissatisfaction/satisfaction. The findings demonstrated that the pre-service teachers’ emotional responses to the streamed lectures tended towards happiness, security, and satisfaction within the typology of affect groups – un/happiness, in/security, and dis/satisfaction. Fewer students reported that the streamed lectures triggered negative feelings of frustration, powerlessness, and inadequacy, and when this occurred, it often pertained to expectations of themselves in the forthcoming field experience in classrooms. Exceptions to this pattern of responses occurred in relation to the fifth streamed lecture presented in a non-interactive slideshow format that compressed a large amount of content. Many students responded to the content of the lecture rather than providing their emotional responses to this lecture, and one student felt “completely disengaged”. The social practice of online writing as blogs enabled the students to articulate their emotions. The findings primarily contribute new understanding about students' wide range of differing emotional states, both positive and negative, experienced in response to streamed live lectures and other learning activities in higher education external coursework. The is important because the majority of previous studies have focused on particular negative emotions, such as anxiety in test taking. The research also highlights the potentials of appraisal theory for studying human emotions in online learning and writing.

Using a video camera to measure the radius of the Earth

A simple but accurate method for measuring the Earth’s radius using a video camera is described. A video camera was used to capture a shadow rising up the wall of a tall building at sunset. A free program called ImageJ was used to measure the time it took the shadow to rise a known distance up the building. The time, distance and length of the sidereal day were used to calculate the radius of the Earth. The radius was measured as 6394.3 +/- 118 km, which is within 1.8% of the accepted average value of 6371 km and well within the experimental error. The experiment is suitable as a high school or university project and should produce a value for Earth’s radius within a few per cent at latitudes towards the equator, where at some times of the year the ecliptic is approximately normal to the horizon.