Unsupervised Broadcast News Video Shot Segmentation and Classification

The relative importance of spatial versus temporal structure in the perception of biological motion: An event-related potential study

Mass media plays an important role during a public health crisis by disseminating information, news and spreading awareness among the masses about combating a health emergency. In the context of the Covid-19 pandemic impacting lives across the globe since its advent in December 2019, this study investigates the role played by TV news channels in India through their unstinted coverage of the same. The study focused on two objectives – 1) finding out viewership trends and consumption patterns among television news audiences in Gujarat during the lockdown and 2) studying the nature of Covid coverage found on mainstream private television news channels in India. A survey with 160 television news viewers was conducted in Ahmedabad and Vadodara. Results demonstrate the reach, significance and impact of television news in a country like India, especially during a health emergency situation. Despite the flourishing of digital content and the rise in internet penetration, television remains a trusted and dependable source of information and knowledge seeking for Indian audiences. In light of the role played by television during the Covid crisis, the future shall witness more scope and opportunities for television to influence viewers with respect to health-related behaviour and practices, development issues and larger global concerns.

Over the years, television news has tended to attract positive appraisal from viewing publics. Most viewers in developed democratic nations place trust in mainstream news broadcasters and, more significantly, believe that they learn a lot from televised news (Gunter, 1987; Towler, 2003; Pew Research, 2012). We might expect therefore that, if people are good judges of their own abilities to learn and have formed well-reasoned opinions about televised news based on their own extensive experiences, tests of learning from news broadcasts should reveal that a lot of information is retained from them. In addition, if we assume that news broadcasters, as trained professionals, know — as they claim to — what their audiences want and need, and know what they are doing in constructing digestible news, then news broadcasts should be produced in a way that will enhance viewers' understanding of current news events and issues. It may therefore come as a surprise that when viewers are tested for what they can remember from a television news programme they recently watched, most of its contents are lost to them. Furthermore, when specific news stories are recalled or recognised afterwards, viewers' understanding of specific details can be confused.KeywordsEpisodic MemoryNews StoryTelevision NewsNews EventNews BroadcastThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

An effective method for speaker identification in news video is presented in this paper, which is based on shot segmentation and exploits both audio and visual cues. Firstly, audio is segmented by shot segmentation based on the observation that there is only one speaker in a shot of news video in most cases. Furthermore, speech/non-speech discrimination is implemented on each shot. Finally, text-independent speaker identification is proposed using audio features on the discriminated speech shots. Experimental results show that our algorithm can obtain satisfactory performance in identifying speakers, so it can be used in real application.

PurposeThe purpose of this paper is to map the variables that affect the customer-based brand equity (CBBE) of media channels (television news) in an emerging market context.Design/methodology/approachThe study adopted structural equation modelling (SEM) to investigate the causal relationships between CBBE and the variables that affect brand equity for television news channels.FindingsThe analysis revealed that localization, ideology, credibility and entertainment are the variables that influence CBBE of television news channels (media brands). Subsequent analysis using SEM indicated that apart from the sole negative impact of entertainment, all the variables had positive impact on brand equity.Research limitations/implicationsThis study is confined to one of the metros of emerging market hence it cannot be generalized. Also the variables that indicate brand equity have been tested only for television news channels hence they may not hold true for other form of television stations.Practical implicationsFor marketers of news channels this study identifies the factors that they need to focus on if they want to garner the equity of the brand in an emerging market scenario.Social implicationsThe content factors identified that influence television news brand equity are reflections of the social requirements of an emerging market. It indicates what the audiences in such markets expect from their television news channels and is part of the social discussion.Originality/valueThe study contributes to brand equity literature by finding the antecedents that can influence any media brand in the emerging market scenario.

Television news is an important medium to convey information to masses. This motivates several stakeholders to monitor and analyze the news broadcasts. Segmentation of streaming broadcast into programs and stories is a necessary first step for such analysis. Television news producers use predefined and unique presentation styles to create the channel content. Presentation styles vary with program and news story category, broadcast time, targeted audience etc. This motivated us to use presentation styles as features for segmenting news broadcasts. We propose a novel approach for characterization of spatio-temporal presentation styles. This involves characterization of spatial styles using a set of (presentation style specific) semantic shot categories derived from LSCOM-Lite Ontology. We also identify features and classifiers to automate the process of shot labeling for spatial style characterization. Further, the temporal presentation styles of shots are modeled using conditional random fields. This spatio-temporal modeling of presentation styles is used for segmenting the broadcast into programs and stories. We have also contributed a 360 hours broadcast video dataset acquired from three Indian English news channels with ground-truth marked semantic shot categories, program genres and story boundaries. Experimentations on this dataset have shown the utility of our proposal for news broadcast video segmentation.

Una investigacion ha analizado distintos tipos de relaciones entre codigos audiovisuales en los generos informativos televisivos de finales de la decada de 2000. La observacion directa de una muestra de noticias y reportajes emitidos en informativos de television generalista y la practica de una serie de entrevistas a profesionales de la informacion audiovisual confirman que las relaciones entre los codigos audiovisuales mas habituales son las que buscan el equilibrio perfecto entre el audio y el video asi como aquellas en las que predomina el codigo linguistico. Ademas se detecta que el transcurso del tiempo influye en la eleccion de un determinado tipo de relacion entre codigos a la hora de elaborar el mensaje audiovisual en noticias y reportajes de distinto tipo y duracion. Un experimento practicado entre sesenta voluntarios demuestra que la audiencia es capaz de identificar tanto los distintos modos de relacionar el audio y el video como su correspondiente grado de coordinacion.

We propose a method for discriminating between a speech shot and a narrated shot to extract genuine speech shots from a broadcast news video. Speech shots in news videos contain a wealth of multimedia information of the speaker, and could thus be considered valuable as archived material. In order to extract speech shots from news videos, there is an approach that uses the position and size of a face region. However, it is difficult to extract them with only such an approach, since news videos contain non-speech shots where the speaker is not the subject that appears in the screen, namely, narrated shots. To solve this problem, we propose a method to discriminate between a speech shot and a narrated shot in two stages. The first stage of the proposed method directly evaluates the inconsistency between a subject and a speaker based on the co-occurrence between lip motion and voice. The second stage of the proposed method evaluates based on the intra- and inter-shot features that focus on the tendency of speech shots. With the combination of both stages, the proposed method accurately discriminates between a speech shot and a narrated shot. In the experiments, the overall accuracy of speech shots extraction by the proposed method was 0.871. Therefore, we confirmed the effectiveness of the proposed method.

The unsupervised representation learning for skeleton-based human action can be utilized in a variety of pose analysis applications. However, previous unsupervised methods focus on modeling the temporal dependencies in sequences, but take less effort in modeling the spatial structure in human action. To this end, we propose a novel unsupervised learning frame-work named Hierarchical Transformer for skeleton-based human action recognition. The Hierarchical Transformer consists of hierarchically aggregated self-attention modules for better capturing the spatial and temporal structure in the skeleton sequences. Furthermore, we propose to predict the motion between adjacent frames as a novel pre-training task for better capturing the long-term dependencies in sequences. Experimental results show that our method outperforms prior state-of-the-art unsupervised methods on NTU RGB+D and NW-UCLA datasets. Besides, our method also achieves state-of-the-art performance when the pre-trained model is transferred to SBU dataset, which demonstrates the generalizability of learned representation.

Television news channels broadcast different kinds of content like debates, interviews, commercials along with news presentations. Real-time detection of these news programs or their retrieval from large volumes of stored broadcast videos is a challenging problem and is a necessary first step for broadcast analytics. News program detection is even harder in Indian context where closed caption text or program markers are not provided by TV news channels (not mandated by law). We propose a two-stage approach to classify news video segments. First, broadcast video shots are classified with multiple labels based on a set of audio-visual features. Second, sequences of these shot features are modeled to detect news programs. Another contribution of this work is the construction of a dataset of 120 hours of shot categories and news programs from Indian English news channels. We have experimented with SVM, HMM and CRF based classifiers and achieved a F1 score of 99% in detecting news programs while experimenting on our dataset.

The widespread of rumors on social media, carrying unreal or even malicious information, brings negative effects on society and individuals, which makes the automatic detection of rumors become particularly important. Most of the previous studies focused on text mining using supervised models based on feature engineering or deep learning models. In recent years, another parallel line of works, which focuses on the spatial structure of message propagation, provides an alternative and promising solution. However, these existing methods in this parallel line largely overlooked the temporal structure information associated with the spatial structure in message propagation. Actually the addition of temporal structure information can make the message propagations be classified from the perspective of spatial–temporal structure, a more fine-grained perspective. Under these observations, this paper proposes a spatial–temporal structure neural network for rumor detection, termed as STS-NN, which treats the spatial structure and the temporal structure as a whole to model the message propagation. All the STS-NN units are parameter shared and consist of three components, including spatial capturer, temporal capturer and integrator, to capture the spatial–temporal information for the message propagation. The results show that our approach obtains better performance than baselines in both rumor classification and early detection.

Segmentation is a low-level processing aimed at the partition of an image in disjoint regions, each one homogeneous with respect to some properties like intensity, texture, shape, etc. Such a task is needed in many high-level processing and applications in such diverse fields as remote-sensing, medical imaging, video coding, and industrial automation, just to name a few. In this work of thesis, we first deal with the problem of color based image segmentation, that is based on some form of homogeneity in the color/spectral properties of the image, and then move to the more complex task of texture based segmentation, where the aim is to recognize complex structures in the image which are typically non homogeneous in terms of spectral properties. Concerning color based segmentation, we focus on a specific class of Markov Random Field (MRF) probability models, namely the Tree-Structured Markov Random Fields (TS-MRF), recently proposed to provide a hierarchical multiscale description of images. Based on such a model, the unsupervised image segmentation is carried out by means of a sequence of nested class splits, where each class is modeled as a local binary MRF. As a first contribution, we first extend the model to generic tree structures removing the unnecessary binary constraint, and then devise a new TS-MRF unsupervised segmentation technique which improves upon the original algorithm by selecting a better tree structure and eliminating spurious classes. Such results are obtained by using the Mean-Shift procedure to estimate the number of pdf modes at each node, and to obtain a more reliable initial clustering for subsequent MRF optimization. To this end, we devise a new reliable and fast clustering algorithm based on the Mean-Shift technique, which makes use of a variable-bandwidth strategy based on the k-Nearest Neighbors (k-NN) technique, and is implemented with a speed-up strategy that cuts significantly the computational complexity. Experimental results on synthetic data and real life images from the remote sensing domain prove the potential of the proposed method. Turning to texture based segmentation, we present a novel multiscale texture model, and a related algorithm for the unsupervised segmentation of color images. Elementary textures are characterized by their spatial interactions with neighboring regions along selected directions. Such interactions are modeled in turn by means of a set of Markov chains, one for each direction, whose parameters are collected in a feature vector that synthetically describes the texture. Based on the feature vectors, the texture are then recursively merged, giving rise to larger and more complex textures, which appear at different scales of observation: accordingly, the model is named Hierarchical Multiple Markov Chain (H-MMC). The Texture Fragmentation and Reconstruction (TFR) algorithm, addresses the unsupervised segmentation problem based on the H-MMC model. The ``fragmentation'' step allows one to find the elementary textures of the model, while the ``reconstruction'' step defines the hierarchical image segmentation based on a probabilistic measure (texture score) which takes into account both region scale and inter-region interactions. The performance of the proposed method was assessed through the Prague segmentation benchmark, based on mosaics of real natural textures, and also tested on real-world natural images from the Berkeley dataset. Finally, an application of the TFR algorithm for the unsupervised segmentation of multiresolution remotely sensed images is proposed, based on the significant presence of textural content in last generation high-resolution satellite images. The obtained results further assess the potential of the proposed method for such real life applications.

The number of speakers is generally greater in broadcast news, therefore speakers frequently occur under varying acoustic conditions. This makes the audio stream contain many speaker turns. Given the small amount of data available (short speech segments), segmentation techniques are not known for their robustness. To address the problems associated with the high number of speaker interventions, we propose an unsupervised speaker segmentation approach that has the advantage of high accuracy of model-selection-based methods. In this approach we use the adapted Gaussian mixture models (GMM) to describe the speaker's characteristics. Experimental results on several journalistic programs show that the proposed method significantly improves segmentation and therefore reduces the diarization error.

The article presents the results of scientific developments concerning the structural organization of sacred landscapes. The methodological basis of the study is the concept of constructive-geographic analysis, which is based on the approaches of the natural and the humanitarian sciences. The system approach to the study of sacred landscapes as a holistic organized territorial structure and a set of methods is used in this work, in particular: structural and logical generalization and system analysis, comparative and geographical, historical and geographical. The author considers the significance of the notion of sacral landscape as being broader than religion per se, and considers it a natural, natural-anthropogenic and anthropogenic system associated with certain symbols of life, myths, significant events, and , indeed ,religious feelings that are of great importance to a person or group of people and requires special respect and protection. The structural organization of all sacred landscapes is characterized by their properties and spatial structure and is closely connected with their social and functional purpose. As a result, such territorial systems can be divided into: confessional, taphal, active, abnormal. The sacred landscape is characterized by polystructurality, that is, the presence of spatial, temporal and morphological structure. In the spatial structure of the sacral landscape, the following components can be distinguished: the sacred object, anthropogenic and technogenic component, the landscape structure and a person with his/ her spiritual experience. In addition, such a structure has a hierarchical construction, where individual, local, regional, national and global levels can be distinguished. This article presents the peculiarities of the temporal structure of sacral landscapes and outlines the external, internal, and the functioning time. Particular attention is paid to the characteristic of internal time, where one can distinguish the following phases of development: the formation of a natural, natural-anthropogenic or anthropogenic landscape; the creation of a spiritual component; loss of sacred human perception of a natural, natural-anthropogenic or anthropogenic landscape; the disappearance of the natural or natural- anthropogenic landscape. Taking into account the morphological structure of the sacred landscape, it is substantiated that religious objects serving as markers of sacred landscapes cannot correspond to one or another morphological unit of the landscape, that is, completely repeat its outlines and boundaries. However, there is a correlation between the type of landscape and the features of the sacred objects that were formed there.

<p>Extreme precipitation events are expected to occur more frequently in a warming climate. Understanding their structure and predicting the exact time and location of precipitation events still remains a challenge because of the high temporal and spatial variability of rainfall. Nationwide weather radar networks are a common tool for investigating precipitation events and their spatial and temporal structure. The German Weather Service (DWD) provides a nationwide climatological radar data set from 2001 to 2020. A reprocessing procedure has been applied to reflectivity measurements in order to obtain precipitation estimates as homogeneous as possible. With an object-oriented analysis, all precipitation events for 11 different durations from 1 to 72 hours exceeding DWD’s official warning level for heavy precipitation have been detected and statistically analysed.</p><p>We will present a comprehensive analysis of all heavy precipitation events that occurred in Germany between 2001 and 2020. We examined their size, duration, location, spatial structure and distribution as well as regional and climatological differences and demonstrate how this information is collected in an online tool for easy access. An assessment of how well these heavy precipitation events were captured by DWD’s network of precipitation stations will be given. Finally, we will present the possibility to use the event detection procedure as an operational tool for assessing and classifying heavy precipitation events and their potential impact in near real-time.</p>