Large Number Of Videos Research Articles

Accelerating temporal action proposal generation via high performance computing

Temporal action recognition always depends on temporal action proposal generation to hypothesize actions and algorithms usually need to process very long video sequences and output the starting and ending times of each potential action in each video suffering from high computation cost. To address this, based on boundary sensitive network we propose a new temporal convolution network called Multipath Temporal ConvNet (MTN), which consists of two parts i.e. Multipath DenseNet and SE-ConvNet. In this work, one novel high performance ring parallel architecture based on Message Passing Interface (MPI) is further introduced into temporal action proposal generation, which is a reliable communication protocol, in order to respond to the requirements of large memory occupation and a large number of videos. Remarkably, the total data transmission is reduced by adding a connection between multiple computing load in the newly developed architecture. It is found that, compared to the traditional Parameter Server architecture, our parallel architecture has higher efficiency on temporal action detection task with multiple GPUs, which is suitable for dealing with the tasks of temporal action proposal generation, especially for large datasets of millions of videos. We conduct experiments on ActivityNet-1.3 and THUMOS14, where our method outperforms other state-of-art temporal action detection methods with high recall and high temporal precision. In addition, a time metric is further proposed here to evaluate the speed performance in the distributed training process.

A Dynamical Generative Model of Social Interactions.

The ability to make accurate social inferences makes humans able to navigate and act in their social environment effortlessly. Converging evidence shows that motion is one of the most informative cues in shaping the perception of social interactions. However, the scarcity of parameterized generative models for the generation of highly-controlled stimuli has slowed down both the identification of the most critical motion features and the understanding of the computational mechanisms underlying their extraction and processing from rich visual inputs. In this work, we introduce a novel generative model for the automatic generation of an arbitrarily large number of videos of socially interacting agents for comprehensive studies of social perception. The proposed framework, validated with three psychophysical experiments, allows generating as many as 15 distinct interaction classes. The model builds on classical dynamical system models of biological navigation and is able to generate visual stimuli that are parametrically controlled and representative of a heterogeneous set of social interaction classes. The proposed method represents thus an important tool for experiments aimed at unveiling the computational mechanisms mediating the perception of social interactions. The ability to generate highly-controlled stimuli makes the model valuable not only to conduct behavioral and neuroimaging studies, but also to develop and validate neural models of social inference, and machine vision systems for the automatic recognition of social interactions. In fact, contrasting human and model responses to a heterogeneous set of highly-controlled stimuli can help to identify critical computational steps in the processing of social interaction stimuli.

The Role of Video in the Flipped Language Classroom

Flipped classrooms have become a widespread form of teaching. Yet, there is no consensus on how to define flipped (language) learning. Several authors consider the use of videos that prepares in-class activities as an essential principle.
 The article presents a study which examined the actual roles of videos in a corpus of 52 descriptions by L2 teachers of flipped language class settings and using Willis’ 1983 framework.
 In the corpus videos played a central role in before-class activities; a large number of videos were used. The roles that videos played in before-class activities in the settings did not all correspond to Willis’ framework; those which did not fit corresponded to direct instruction. The definition of a flipped setting was found to be unclear, as in a quarter of the descriptions the criteria did not apply. Video was not found to be necessarily constitutive for flipped language classes.

Exploiting Audio-Visual Consistency with Partial Supervision for Spatial Audio Generation

Human perceives rich auditory experience with distinct sound heard by ears. Videos recorded with binaural audio particular simulate how human receives ambient sound. However, a large number of videos are with monaural audio only, which would degrade the user experience due to the lack of ambient information. To address this issue, we propose an audio spatialization framework to convert a monaural video into a binaural one exploiting the relationship across audio and visual components. By preserving the left-right consistency in both audio and visual modalities, our learning strategy can be viewed as a self-supervised learning technique, and alleviates the dependency on a large amount of video data with ground truth binaural audio data during training. Experiments on benchmark datasets confirm the effectiveness of our proposed framework in both semi-supervised and fully supervised scenarios, with ablation studies and visualization further support the use of our model for audio spatialization.

Overview of TikTok's most viewed dermatologic content and assessment of its reliability

Overview of TikTok's most viewed dermatologic content and assessment of its reliability

Vehicle theft recognition from surveillance video based on spatiotemporal attention

Frequent vehicle thefts have a highly detrimental impact on public safety. Thanks to surveillance equipment distributed throughout a city, a large number of videos that can be used to recognize vehicle theft are available. However, vehicle theft behavior has the characteristics of a small criminal target and small movement. Hence, the existing action recognition algorithms cannot be directly applied for the recognition of vehicle theft. In this paper, we propose a method for vehicle theft recognition based on a spatiotemporal attention mechanism. First, a database of vehicle theft is established by collecting videos from the Internet and an existing dataset. Then, we establish a vehicle theft recognition network and introduce a spatiotemporal attention mechanism for application when extracting the spatiotemporal features of theft. Through the learning of adaptive feature weights, the features that contribute most greatly to recognition are emphasized. Simulation experiments show that our proposed algorithm can achieve 97.04% accuracy on the collected vehicle theft database.

A Cross-Modal Learning Approach for Recognizing Human Actions

Since a large number of videos have sprung up, the human-centric visual understanding tasks especially human action recognition (HAR) is under great demand. The key issue of the HAR is to capture spatial-temporal collaborative representations by mining abstract and discriminative high-level features. In this article, we propose a cross-modal learning framework that mainly includes an alignment net and a fusion net to improve the performance of the HAR. First, extracted different modal information is mapped into a common subspace to align, which compensates for the spatial-temporal discrepancies. Then, the aligned features are further fused to generate complementary, correlated, and consistent representations. Finally, the learnt features are input to the classifier for recognition. The experimental results have shown that our proposed approach can outperform several state-of-the-art baseline approaches.

NARRATIVAS HISTÓRICAS EM DISPUTA: UM ESTUDO DE CASO NO YOUTUBE

RESUMO Este artigo pretende tratar das diferentes narrativas que envolvem a disciplina da história na plataforma YouTube. A partir de um mapeamento, foi analisada uma série de vídeos, e de uma análise quantitativa e qualitativa desses materiais foi possível inferir, por um lado, uma inexpressividade das produções feitas por instituições públicas ou universitárias e, por outro, um grande número de vídeos que trazem uma visão revisionista ou contestadora da história acadêmica. Concluiu-se, a partir dessas investigações, a respeito da necessidade de uma maior atuação da academia e dos historiadores profissionais nesses espaços de produção e de discussão de narrativas, o que poderia responder a uma demanda educacional identificada que ocorre tanto dentro do espaço físico e formal da sala de aula quanto fora dele, em espaços online que, visando a se constituir em oposição à escola formal, oferecem-se como alternativas nas disputas das narrativas da história.

Theoretical frameworks for constructing a model for analysis of Islamic discourse on YouTube: one view from Serbia

The starting point of this paper an assumption that videos on the YouTube Internet portal relating to Islam produce intolerant and radical discourse that may be a factor of radicalization through their contents, the ways of the shaping of the content, and by their presentation modes. Thus, the aim of the study is to present theoretical frameworks and possible research tools needed to construct a model for the analysis of the Islamic discourse on YouTube. This paper draws on various theoretical approaches, indicate a large number of potential analytical concepts, sub-models, instruments and criteria. It argues in favour of theoretical pluralism, as well as a pluralistic approach and models, openness to accepting diverse methodological instruments, and stresses the need of multidisciplinary in the analysis. Consideration of the relationship between the concepts of politics, discourse and Islamic discourse led to a conclusion about the need of a particular emphasis on the use of critical discourse analysis (CDA) and political discourse analysis (PDA) settings. The results of the research indicate that Islamic discourse on YouTube in a large number of videos, with more or less open manipulative content (visible to the average viewer), aims at consolidating an unquestionable framework of understanding and knowledge of reality – as well as comprehending of the history. Designing a framework for understanding reality and through associated eschatology, the rules of everyday action, principles and limits of tolerance, and identity – especially through discursive strategies and argumentation and beliefs that establish enemies – Islamic discourse on YouTube interprets reality, constructs and impose the only possible reality, and the necessity of achieving the ideal universal Islamic perfection.

Video Summarization Using Highlight Detection and Pairwise Deep Ranking Model

With mobile phones and camera enabled devices becoming pervasive and user-friendly, a large number of videos are being shot every day and uploaded to social media and video streaming websites. This makes them an important information dispensing tool. Searching and analysing such large amount of videos is an extremely tedious task. Thus, automatic video summarization is used to produce a short, informative summary of a long video, which is useful in the indexing and the classification of such videos in the video database. Video summarization is a very challenging task. This work aims to generate a video summary, by modelling a two stream architecture consisting of deep convolutional neural network in each stream for extracting both spatial and temporal information of a video. Two dimensional Convolutional Neural Network (2D CNN) is used to exploit spatial information whereas a three dimensional Convolutional Neural Network (3D CNN) is used to exploit temporal information to generate highlight scores for segments of the video. The segment scores from each stream are fused to detect highlight segments from the video. Furthermore, as the highlight result depicts only a relative level of interest of a user in a video, the DCNN in each stream is trained with a pairwise deep ranking model. The goal is making the highlight score of highlight segment higher than non-highlight segment by optimizing the model. The obtained highlight segments are then used to summarize a video.

The adaptive key frames extraction method based on representativeness and independence features

For the existing traditional method to extract key frames from the video sequence with gradual change of content, there are problems of having too much redundancy or too much loss in the extraction results and other issues, a new key frames extracting method for adaptively extracting the video key frames has proposed in the paper. The experimental results have shown that the proposed method can effectively extract video key frames, especially for the video with gradual change of content, which can express more succinctly and comprehensively. The experiments on a large number of videos containing various scenes have shown that the key frames extracted by the proposed algorithm can more fully express the main content of video. When dealing with content gradient video, the effect is better. The content adaptively determines the number of key frames based on video.

The adaptive key frames extraction method based on representativeness and independence features

For the existing traditional method to extract key frames from the video sequence with gradual change of content, there are problems of having too much redundancy or too much loss in the extraction results and other issues, a new key frames extracting method for adaptively extracting the video key frames has proposed in the paper. The experimental results have shown that the proposed method can effectively extract video key frames, especially for the video with gradual change of content, which can express more succinctly and comprehensively. The experiments on a large number of videos containing various scenes have shown that the key frames extracted by the proposed algorithm can more fully express the main content of video. When dealing with content gradient video, the effect is better. The content adaptively determines the number of key frames based on video.

EVALUATION OF DEBRIS FLOW DYNAMIC CHARACTERISTICS BY VIDEO MATERIALS

For mud dams construction it is necessary to clarify characteristics of debris flow such as flow velocity and pressure on the barrier. Determining these characteristics often causes difficulties due to the fact that debris flow is rather rare event and constant monitoring of them is carried out at mud flow observation station stations, which are few worldwide. Currently large number of videos have appeared on the Internet that captures debris flow descent. This material can be used to obtain not only qualitative, but also quantitative characteristics of the debris flow. In cases, when it is possible to determine the scale and specific location of debris flow on a video, measuring its velocity and other characteristics is a workable task. This research present an attempt to quantify the debris flow velocity based on the video materials with the subsequent comparison of the results obtained using various methods. The object of our study was the debris flow that came down in Austria, in Firgen on August 4, 2012. The survey was carried out from several angles, which made it possible to select a section of the channel to measure the velocity of debris flow wave train. Calculation of flow velocity and pressure on the barrier was conducted by several methods developed by various researchers. Debris flow velocities measured on the video are minimum – 7,4 m/s, maximum - 10 m/s. Differences between calculated by various methods and measured on the video values range from 0,1 m/s to 4,8 m/s.

The usefulness of YouTube™ videos as a source of information on burning mouth syndrome.

The aim of this review was to investigate quality assessment and quantitative information on burning mouth syndrome (BMS) on YouTube™ videos. An electronic search of YouTube™ videos on BMS was performed using subject headings, keywords and synonyms of BMS. For each video, quality information, via the quality assessment score (QAS) and quantitative information, such as the total number of results per term searched, date of upload, clip length, view count, source category (professional, personal, educational, mixed), number of likes and dislikes and YouTube™ category were evaluated. The search identified 3768 videos of which only 114 were included in our analysis: the quality of videos was very poor (mean±standard deviation=1.76±1.64). The most representative categories were "educational" with 53 (46.5%) videos scoring between 0 and 4 and "mixed" with 24 (21.1%) videos scoring between 0 and 2. Significant differences were found among all four categories in relation to length in seconds (P<0.001), number of views (P=0.006) and interaction index (P=0.001), as well as between "professional" and the other categories and between "educational" and "personal" in relation to length in seconds (P<0.001), between "educational" and "personal" (P<0.001) and "personal" and "mixed" category in relation to interaction index (P=0.003) and between "professional" and "personal" in relation to views per day (P=0.015). A large number of videos regarding BMS have published on YouTube™, with a wide range of distribution regarding their length and views. Unfortunately, they presented unreliable information which is most often scientifically inaccurate. Health care institutions and professionals should be more actively involved in improving high-quality e-information about BMS on YouTube™.

Front-Vehicle Detection in Video Images Based on Temporal and Spatial Characteristics.

Assisted driving and unmanned driving have been areas of focus for both industry and academia. Front-vehicle detection technology, a key component of both types of driving, has also attracted great interest from researchers. In this paper, to achieve front-vehicle detection in unmanned or assisted driving, a vision-based, efficient, and fast front-vehicle detection method based on the spatial and temporal characteristics of the front vehicle is proposed. First, a method to extract the motion vector of the front vehicle is put forward based on Oriented FAST and Rotated BRIEF (ORB) and the spatial position constraint. Then, by analyzing the differences between the motion vectors of the vehicle and those of the background, feature points of the vehicle are extracted. Finally, a feature-point clustering method based on a combination of temporal and spatial characteristics are applied to realize front-vehicle detection. The effectiveness of the proposed algorithm is verified using a large number of videos.

Assessment of YouTube Videos on H Index

H index is a popular measure. YouTube provides free of cost video content on different topics. Anyone can upload the videos on YouTube after setting up an account without any verification by experts. Large numbers of videos have been uploaded on YouTube which may or may not be relevant. This study is aimed at evaluating the usefulness of YouTube videos on h-index. YouTube (www.youtube.com) was searched with the search term “h index” on 18th August 2018. The search results were filtered by video type. The first 20 videos were selected for assessment, but after screening the four non-English videos were excluded. These videos have been watched and evaluated based on six aspects. The study found that no video covered all six aspects. 25% of the screened videos were found very useful, 38% of the videos were found slightly useful, 31% of the videos were found moderately useful and only 1 video was found not useful. It was also found that majority of the videos covered definition and example and very few videos discussed advantages of h-index.

Comparison of Debris Flow Characteristics Obtained by Using Video Materials and Modeling

Velocity of debris flow is one of the most important characteristics for the protective construction design. Since debris flows are rare events, and observations are conducted only on stations in Russia, Ukraine, Italy, Switzerland, USA, China, Japan and New Zealand, the velocity is calculated rather than measured. Nowadays, a large number of videos with passing debris flows have appeared on the Internet. Scientists can use such video materials to obtain qualitative and quantitative characteristics of the debris flow. Therefore, the aim of our research is an attempt to measure the debris flow velocity using video materials and compare the obtained results with the calculated values using various methods. The debris flow that came down in Firgen, Austria on August 4, 2012 was chosen as the object of our study. The video was carried out from several angles, so it was possible to select a section of the channel, through which we could measure the debris flows waves velocity. In addition, we calculated the velocities of waves by formulas adopted in the regulatory documents and compared with the measured by video values. During the video analysis, debris flow velocities at different sites were observed: minimum—7.2 m/s and maximum—10 m/s. The calculated values varied from 4.5 m/s to 11.4 m/s. Moreover, we applied model of the transport-shear process of debris flow formation developed by Yu. B. Vinogradov. When we were comparing the obtained debris flow discharges with results from Austrian colleagues, we found out that the values were similar to each other. However, internal scatter in the model changed from 151 to 190 m3/s, while in the report of Austrian colleagues the discharges were from 80 to 250 m3/s.

Multimodal Sentiment Analysis: A Comparison Study

Sentiments and emotions play a pivotal role in our daily lives. They assist decision making, learning, communication and situation awareness in human environments. Sentiment analysis is mainly focused on the automatic recognition of opinions’ polarity, as positive or negative. Nowadays, sentiment analysis is replacing the old web based survey and traditional survey methods that conducted by deferent companies for finding public opinion about entities like products and services in order to improve their marketing strategy and product of advertisement, at the same time sentiment analysis improves customer service. Large number of videos is being uploaded online every day. Video files contain text, visual and audio features that complement each other. Multimodality is defined by analyzing more than one modality, Multimodal Sentiment Analysis refers to the combination of two or more input models in order to improve the performance of the analysis; a combination of text and audio-visual inputs is an example. The automatic analysis of multimodal opinion involves a deep understanding of natural languages, audio and video processing, whereas researchers are continuing to improve them. This paper focuses on multimodal sentiment analysis as text, audio and video, by giving a complete image of it and related dataset available and providing brief details for each type, in addition to that present the recent trend of researches in the multimodal sentiment analysis and its related fields will be explored.

All play and no work: skits and models in teaching skeletal muscle physiology.

All play and no work: skits and models in teaching skeletal muscle physiology.

Video logo removal detection based on sparse representation

With the popularity of multimedia editing tools, more and more forged multimedia content appeared on the network. Thus, the legal authorities need novel techniques to distinguish copyright infringements from a large number of videos on the Internet. Since logo removal is a common editing operation during unauthorized reproduction, logo removal detection is often equivalent to copyright infringements detection to some extent. In this paper, we proposed a video forensics framework for logo removal detection. Our framework mainly contains two stages: the removal traces detection and the removal region location. In the first stage, we use sparse representation to show the difference between the tampered areas and the original areas in sparsity. In the second stage, spatial priors and temporal correlations are used to refine the location of the removal regions. Finally, a spatiotemporal suspected region can obviously show the edited regions. The proposed method is validated on our video logo removal dataset by extensive experiments, showing promising results.