自注意力机制的视频摘要模型

针对如何高效地识别出视频中具有代表性的内容问题,提出了一种对不同的视频帧赋予不同重要性的视频摘要算法.首先使用长短期记忆网络来建模视频序列的时序关系,然后利用自注意力机制建模视频中不同帧的重要性程度并提取全局特征,最后通过每一帧回归得到的重要性得分进行采样,并使用强化学习策略优化模型参数.其中,强化学习的动作定义为每一帧选或者不选,状态定义为当前这个视频的选择情况,反馈信号使用多样性和代表性代价.在2个公开数据集SumMe和TVSum中进行视频摘要实验,并使用F-度量来衡量这2个数据集上不同视频摘要算法的准确度,实验结果表明,提出的视频摘要算法结果要优于其他算法.     

Dynamic video summarization using two-level redundancy detection

The mushroom growth of video information, consequently, necessitates the progress of content-based video analysis techniques. Video summarization, aiming to provide a short video summary of the original video document, has drawn much attention these years. In this paper, we propose an algorithm for video summarization with a two-level redundancy detection procedure. By video segmentation and cast indexing, the algorithm first constructs story boards to let users know main scenes and cast (when this is a video with cast) in the video. Then it removes redundant video content using hierarchical agglomerative clustering in the key frame level. The impact factors of scenes and key frames are defined, and parts of key frames are selected to generate the initial video summary. Finally, a repetitive frame segment detection procedure is designed to remove redundant information in the initial video summary. Results of experimental applications on TV series, movies and cartoons are given to illustrate the proposed algorithm.

基于改进的双向长短期记忆网络的视频摘要生成模型

针对传统视频摘要方法往往没有考虑时序信息以及提取的视频特征过于复杂、易出现过拟合现象的问题,提出一种基于改进的双向长短期记忆（BiLSTM）网络的视频摘要生成模型。首先,通过卷积神经网络（CNN）提取视频帧的深度特征,而且为了使生成的视频摘要更具多样性,采用BiLSTM网络将深度特征识别任务转换为视频帧的时序特征标注任务,让模型获得更多上下文信息;其次,考虑到生成的视频摘要应当具有代表性,因此通过融合最大池化在降低特征维度的同时突出关键信息以淡化冗余信息,使模型能够学习具有代表性的特征,而特征维度的降低也减少了全连接层需要的参数,避免了过拟合问题;最后,预测视频帧的重要性分数并转换为镜头分数,以此选取关键镜头生成视频摘要。实验结果表明,在标准数据集TvSum和SumMe上,改进后的视频摘要生成模型能提升生成视频摘要的准确性;而且它的F1-score值也比基于长短期记忆（LSTM）网络的视频摘要模型DPPLSTM在两个数据集上分别提高1.4和0.3个百分点。     

基于空时变换网络的视频摘要生成

视频摘要生成是计算机视觉领域必不可少的关键任务,这一任务的目标是通过选择视频内容中信息最丰富的部分来生成一段简洁又完整的视频摘要,从而对视频内容进行总结.所生成的视频摘要通常为一组有代表性的视频帧（如视频关键帧）或按时间顺序将关键视频片段缝合所形成的一个较短的视频.虽然视频摘要生成方法的研究已经取得了相当大的进展,但现有的方法存在缺乏时序信息和特征表示不完备的问题,很容易影响视频摘要的正确性和完整性.为了解决视频摘要生成问题,本文提出一种空时变换网络模型,该模型包括三大模块,分别为：嵌入层、特征变换与融合层、输出层.其中,嵌入层可同时嵌入空间特征和时序特征,特征变换与融合层可实现多模态特征的变换和融合,最后输出层通过分段预测和关键镜头选择完成视频摘要的生成.通过空间特征和时序特征的分别嵌入,以弥补现有模型对时序信息表示的不足;通过多模态特征的变换和融合,以解决特征表示不完备的问题.我们在两个基准数据集上做了充分的实验和分析,验证了我们模型的有效性.     

A video summarization approach based on the emulation of bottom-up mechanisms of visual attention

This work addresses the development of a computational model of visual attention to perform the automatic summarization of digital videos from television archives. Although the television system represents one of the most fascinating media phenomena ever created, we still observe the absence of effective solutions for content-based information retrieval from video recordings of programs produced by this media universe. This fact relates to the high complexity of the content-based video retrieval problem, which involves several challenges, among which we may highlight the usual demand on video summaries to facilitate indexing, browsing and retrieval operations. To achieve this goal, we propose a new computational visual attention model, inspired on the human visual system and based on computer vision methods (face detection, motion estimation and saliency map computation), to estimate static video abstracts, that is, collections of salient images or key frames extracted from the original videos. Experimental results with videos from the Open Video Project show that our approach represents an effective solution to the problem of automatic video summarization, producing video summaries with similar quality to the ground-truth manually created by a group of 50 users.     

一种层次的电影视频摘要生成方法

合理地组织视频数据对于基于内容的视频分析和检索有着重要的意义。提出了一种基于运动注意力模型的电影视频摘要生成方法。首先给出了一种基于滑动镜头窗的聚类算法将相似的镜头组织成为镜头类;然后根据电影视频场景内容的发展模式,在定义两个镜头类的3种时序关系的基础上,提出了一种基于镜头类之间的时空约束关系的场景检测方法;最后利用运动注意力模型选择场景中的重要镜头和代表帧,由选择的代表帧集合和重要镜头的关键帧集合建立层次视频摘要(场景级和镜头级)。该方法较全面地涵盖了视频内容,又突出了视频中的重要内容,能够很好地应用于电影视频的快速浏览和检索。     

Hierarchical Indexing Structure for Efficient Similarity Search in Video Retrieval

With the rapid increase in both centralized video archives and distributed WWW video resources, content-based video retrieval is gaining its importance. To support such applications efficiently, content-based video indexing must be addressed. Typically, each video is represented by a sequence of frames. Due to the high dimensionality of frame representation and the large number of frames, video indexing introduces an additional degree of complexity. In this paper, we address the problem of content-based video indexing and propose an efficient solution, called the ordered VA-file (OVA-file) based on the VA-file. OVA-file is a hierarchical structure and has two novel features: 1) partitioning the whole file into slices such that only a small number of slices are accessed and checked during k nearest neighbor (kNN) search and 2) efficient handling of insertions of new vectors into the OVA-file, such that the average distance between the new vectors and those approximations near that position is minimized. To facilitate a search, we present an efficient approximate kNN algorithm named ordered VA-LOW (OVA-LOW) based on the proposed OVA-file. OVA-LOW first chooses possible OVA-slices by ranking the distances between their corresponding centers and the query vector, and then visits all approximations in the selected OVA-slices to work out approximate kNN. The number of possible OVA-slices is controlled by a user-defined parameter delta. By adjusting delta, OVA-LOW provides a trade-off between the query cost and the result quality. Query by video clip consisting of multiple frames is also discussed. Extensive experimental studies using real video data sets were conducted and the results showed that our methods can yield a significant speed-up over an existing VA-file-based method and (distance with high query result quality. Furthermore, by incorporating temporal correlation of video content, our methods achieved much more efficient performance     

AVCD-FRA: A novel solution to automatic video cut detection using fuzzy-rule-based approach

Video shot boundary detection (SBD) is a fundamental step in automatic video content analysis toward video indexing, summarization and retrieval. Despite the beneficial previous works in the literature, reliable detection of video shots is still a challenging issue with many unsolved problems. In this paper, we focus on the problem of hard cut detection and propose an automatic algorithm in order to accurately determine abrupt transitions from video. We suggest a fuzzy rule-based scene cut identification approach in which a set of fuzzy rules are evaluated to detect cuts. The main advantage of the proposed method is that, we incorporate spatial and temporal features to describe video frames, and model cut situations according to temporal dependency of video frames as a set of fuzzy rules. Also, while existing cut detection algorithms are mainly threshold dependent; our method identifies cut transitions using a fuzzy logic which is more flexible. The proposed algorithm is evaluated on a variety of video sequences from different genres. Experimental results, in comparison with the most standard cut detection algorithms confirm our method is more robust to object and camera movements as well as illumination changes.     

Spatial-temporal joint probability images for video segmentation

Effective annotation and content-based search for videos in a digital library require a preprocessing step of detecting, locating and classifying scene transitions, i.e., temporal video segmentation. This paper proposes a novel approach—spatial-temporal joint probability image (ST-JPI) analysis for temporal video segmentation. A joint probability image (JPI) is derived from the joint probabilities of intensity values of corresponding points in two images. The ST-JPT, which is a series of JPIs derived from consecutive video frames, presents the evolution of the intensity joint probabilities in a video. The evolution in a ST-JPI during various transitions falls into one of several well-defined linear patterns. Based on the patterns in a ST-JPI, our algorithm detects and classifies video transitions effectively.Our study shows that temporal video segmentation based on ST-JPIs is distinguished from previous methods in the following way: (1) It is effective and relatively robust not only for video cuts but also for gradual transitions; (2) It classifies transitions on the basis of predefined evolution patterns of ST-JPIs during transitions; (3) It is efficient, scalable and suitable for real-time video segmentation. Theoretical analysis and experimental results of our method are presented to illustrate its efficacy and efficiency.     

Unsupervised Video Hashing via Deep Neural Network

Hashing is a common solution for content-based multimedia retrieval by encoding high-dimensional feature vectors into short binary codes. Previous works mainly focus on image hashing problem. However, these methods can not be directly used for video hashing, as videos contain not only spatial structure within each frame, but also temporal correlation between successive frames. Several researchers proposed to handle this by encoding the extracted key frames, but these frame-based methods are time-consuming in real applications. Other researchers proposed to characterize the video by averaging the spatial features of frames and then the existing hashing methods can be adopted. Unfortunately, the sort of “video” features does not take the correlation between frames into consideration and may lead to the loss of the temporal information. Therefore, in this paper, we propose a novel unsupervised video hashing framework via deep neural network, which performs video hashing by incorporating the temporal structure as well as the conventional spatial structure. Specially, the spatial features of videos are obtained by utilizing convolutional neural network, and the temporal features are established via long-short term memory. After that, the time series pooling strategy is employed to obtain the single feature vector for each video. The obtained spatio-temporal feature can be applied to many existing unsupervised hashing methods. Experimental results on two real datasets indicate that by employing the spatio-temporal features, our hashing method significantly improves the performance of existing methods which only deploy the spatial features, and meanwhile obtains higher mean average precision compared with the state-of-the-art video hashing methods.     

Video flickering removal using temporal reconstruction optimization

In this paper, we introduce an approach to remove the flickers in the videos, and the flickers are caused by applying image-based processing methods to original videos frame by frame. First, we propose a multi-frame based video flicker removal method. We utilize multiple temporally corresponding frames to reconstruct the flickering frame. Compared with traditional methods, which reconstruct the flickering frame just from an adjacent frame, reconstruction with multiple temporally corresponding frames reduces the warp inaccuracy. Then, we optimize our video flickering method from following aspects. On the one hand, we detect the flickering frames in the video sequence with temporal consistency metrics, and just reconstructing the flickering frames can accelerate the algorithm greatly. On the other hand, we just choose the previous temporally corresponding frames to reconstruct the output frames. We also accelerate our video flicker removal with GPU. Qualitative experimental results demonstrate the efficiency of our proposed video flicker method. With algorithmic optimization and GPU acceleration, the time complexity of our method also outperforms traditional video temporal coherence methods.

     

VSUMM: A mechanism designed to produce static video summaries and a novel evaluation method

The fast evolution of digital video has brought many new multimedia applications and, as a consequence, has increased the amount of research into new technologies that aim at improving the effectiveness and efficiency of video acquisition, archiving, cataloging and indexing, as well as increasing the usability of stored videos. Among possible research areas, video summarization is an important topic that potentially enables faster browsing of large video collections and also more efficient content indexing and access. Essentially, this research area consists of automatically generating a short summary of a video, which can either be a static summary or a dynamic summary. In this paper, we present VSUMM, a methodology for the production of static video summaries. The method is based on color feature extraction from video frames and k-means clustering algorithm. As an additional contribution, we also develop a novel approach for the evaluation of video static summaries. In this evaluation methodology, video summaries are manually created by users. Then, several user-created summaries are compared both to our approach and also to a number of different techniques in the literature. Experimental results show - with a confidence level of 98% - that the proposed solution provided static video summaries with superior quality relative to the approaches to which it was compared.     

Accessing Video Contents through Key Objects over IP

In order to support content-based video database access over the Internet Protocol (IP), achieving the following objectives are important: (i) video query by a representative object (key object) or some statistical characterization of the target contents, (ii) bandwidth-efficient browsing over IP, and (iii) scalable and user-centric video transmission over a heterogeneous and variable-bandwidth network. We present a video object extraction and scalable coding system designed to meet the above objectives. In our system, key objects of meaning to video database users are generated via a human-computer-interaction procedure, and are tracked across frames. Given a key object, an algorithm classifies a subset of its VOPs as key VOPs. This subset forms the basis of a highly bandwidth-efficient base layer for supporting activities such as browsing and refining queries. Over the base layer, a number of enhancement layers can be defined to progressively increase the spatial and temporal resolutions of retrieved video. It is expected that heterogeneous users can subscribe to different numbers of the enhancement layers according to their own conditions, such as access authorization, available connection bandwidth, and quality preference.     

Video dehazing with spatial and temporal coherence

This paper describes a new framework for video dehazing, the process of restoring the visibility of the videos taken under foggy scenes. The framework builds upon techniques in single image dehazing, optical flow estimation and Markov random field. It aims at improving the temporal and spatial coherence of the dehazed video. In this framework, we first extract the transmission map frame-by-frame using guided filter, then estimate the forward and backward optical flow between two neighboring frames to find the matched pixels. The flow fields are used to help us building an MRF model on the transmission map to improve the spatial and temporal coherence of the transmission. The proposed algorithm is verified in both real and synthetic videos. The results demonstrate that our algorithm can preserve the spatial and temporal coherence well. With more coherent transmission map, we get better refocusing effect. We also apply our framework on improving the video coherence on the application of video denoising.     

Automatic video data structuring through shot partitioning and key-frame computing

In video processing, a common first step is to segment the videos into physical units, generally called shots. A shot is a video segment that consists of one continuous action. In general, these physical units need to be clustered to form more semantically significant units, such as scenes, sequences, programs, etc. This is the so-called story-based video structuring. Automatic video structuring is of great importance for video browsing and retrieval. The shots or scenes are usually described by one or several representative frames, called key-frames. Viewed from a higher level, key frames of some shots might be redundant in terms of semantics. In this paper, we propose automatic solutions to the problems of: (i) video partitioning, (ii) key frame computing, (iii) key frame pruning. For the first problem, an algorithm called “net comparison” is devised. It is accurate and fast because it uses both statistical and spatial information in an image and does not have to process the entire image. For the last two problems, we develop an original image similarity criterion, which considers both spatial layout and detail content in an image. For this purpose, coefficients of wavelet decomposition are used to derive parameter vectors accounting for the above two aspects. The parameters exhibit (quasi-) invariant properties, thus making the algorithm robust for many types of object/camera motions and scaling variances. The novel “seek and spread” strategy used in key frame computing allows us to obtain a large representative range for the key frames. Inter-shot redundancy of the key-frames is suppressed using the same image similarity measure. Experimental results demonstrate the effectiveness and efficiency of our techniques.     

基于仿射传播聚类的视频关键帧提取方法

关键帧提取是基于内容的视频摘要生成中的一个重要技术.首次引入仿射传播聚类方法来提取视频关键帧.该方法结合两个连续图像帧的颜色直方图交,通过消息传递,实现数据点的自动聚类.并与k means和SVC(support vector clustering)算法的关键帧提取方法进行了比较.实验结果表明,AP(Affinity Propagation)聚类的关键帧提取速度快,准确性高,生成的视频摘要具有良好的压缩率和内容涵盖率.     

A scalable signature scheme for video authentication

This paper addresses the problem of ensuring the integrity of a digital video and presents a scalable signature scheme for video authentication based on cryptographic secret sharing. The proposed method detects spatial cropping and temporal jittering in a video, yet is robust against frame dropping in the streaming video scenario. In our scheme, the authentication signature is compact and independent of the size of the video. Given a video, we identify the key frames based on differential energy between the frames. Considering video frames as shares, we compute the corresponding secret at three hierarchical levels. The master secret is used as digital signature to authenticate the video. The proposed signature scheme is scalable to three hierarchical levels of signature computation based on the needs of different scenarios. We provide extensive experimental results to show the utility of our technique in three different scenarios—streaming video, video identification and face tampering.