In this paper, we introduce Coarse-Fine Networks, a two-stream architecture which benefits from different abstractions of temporal resolution to learn better video representations for long-term motion. Traditional Video models process inputs at one (or few) fixed temporal resolution without any dynamic frame selection. However, we argue that, processing multiple temporal resolutions of the input and doing so dynamically by learning to estimate the importance of each frame can largely improve video representations, specially in the domain of temporal activity localization. To this end, we propose (1) Grid Pool, a learned temporal downsampling layer to extract coarse features, and, (2) Multi-stage Fusion, a spatio-temporal attention mechanism to fuse a fine-grained context with the coarse features. We show that our method outperforms the state-of-the-arts for action detection in public datasets including Charades with a significantly reduced compute and memory footprint. The code is available at https://github.com/kkahatapitiya/Coarse-Fine-Networks
Salesforce Research（Salesforce研究院）；Fudan University（复旦大学）；University of Maryland（马里兰大学）
This paper presents LiteEval, a simple yet effective coarse-to-fine framework for resource efficient video recognition, suitable for both online and offline scenarios. Exploiting decent yet computationally efficient features derived at a coarse scale with a lightweight CNN model, LiteEval dynamically decides on-the-fly whether to compute more powerful features for incoming video frames at a finer scale to obtain more details. This is achieved by a coarse LSTM and a fine LSTM operating cooperatively, as well as a conditional gating module to learn when to allocate more computation. Extensive ex…
Beijing University of Posts and Telecommunications（北京邮电大学）；Tencent PCG（腾讯平台与内容事业群）；Tianjin University（天津大学）
The current state-of-the-art methods for video corpus moment retrieval (VCMR) often use similarity-based feature alignment approach for the sake of convenience and speed. However, late fusion methods like cosine similarity alignment are unable to make full use of the information from both query texts and videos. In this paper, we combine feature alignment with feature fusion to promote the performance on VCMR.
In this paper, we introduce a challenging new dataset, MLB-YouTube, designed for fine-grained activity detection. The dataset contains two settings: segmented video classification as well as activity detection in continuous videos. We experimentally compare various recognition approaches capturing temporal structure in activity videos, by classifying segmented videos and extending those approaches to continuous videos. We also compare models on the extremely difficult task of predicting pitch speed and pitch type from broadcast baseball videos. We find that learning temporal structure is valua…
Videos can be manipulated by duplicating a sequence of consecutive frames with the goal of concealing or imitating a specific content in the same video. In this paper, we propose a novel coarse-to-fine framework based on deep Convolutional Neural Networks to automatically detect and localize such frame duplication. First, an I3D network finds coarse-level matches between candidate duplicated frame sequences and the corresponding selected original frame sequences. Then a Siamese network based on ResNet architecture identifies fine-level correspondences between an individual duplicated frame and…
The University of Hong Kong（香港大学）；Alibaba Group（阿里巴巴集团）；The Hong Kong Polytechnic University（香港理工大学）
High dynamic range (HDR) video reconstruction from sequences captured with alternating exposures is a very challenging problem. Existing methods often align low dynamic range (LDR) input sequence in the image space using optical flow, and then merge the aligned images to produce HDR output. However, accurate alignment and fusion in the image space are difficult due to the missing details in the over-exposed regions and noise in the under-exposed regions, resulting in unpleasing ghosting artifacts. To enable more accurate alignment and HDR fusion, we introduce a coarse-to-fine deep learning fra…
In problems such as sports video analytics, it is difficult to obtain accurate frame level annotations and exact event duration because of the lengthy videos and sheer volume of video data. This issue is even more pronounced in fast-paced sports such as ice hockey. Obtaining annotations on a coarse scale can be much more practical and time efficient. We propose the task of event detection in coarsely annotated videos. We introduce a multi-tower temporal convolutional network architecture for the proposed task. The network, with the help of multiple receptive fields, processes information at va…
Current state-of-the-art human activity recognition is focused on the classification of temporally trimmed videos in which only one action occurs per frame. We propose a simple, yet effective, method for the temporal detection of activities in temporally untrimmed videos with the help of untrimmed classification. Firstly, our model predicts the top k labels for each untrimmed video by analysing global video-level features. Secondly, frame-level binary classification is combined with dynamic programming to generate the temporally trimmed activity proposals. Finally, each proposal is assigned a …
Video denoising aims to recover high-quality frames from the noisy video. While most existing approaches adopt convolutional neural networks~(CNNs) to separate the noise from the original visual content, however, CNNs focus on local information and ignore the interactions between long-range regions in the frame. Furthermore, most related works directly take the output after basic spatio-temporal denoising as the final result, leading to neglect the fine-grained denoising process. In this paper, we propose a Dual-stage Spatial-Channel Transformer for coarse-to-fine video denoising, which inheri…
We propose an efficient approach for activity detection in video that unifies activity categorization with space-time localization. The main idea is to pose activity detection as a maximum-weight connected subgraph problem. Offline, we learn a binary classifier for an activity category using positive video exemplars that are "trimmed" in time to the activity of interest. Then, given a novel \emph{untrimmed} video sequence, we decompose it into a 3D array of space-time nodes, which are weighted based on the extent to which their component features support the learned activity model. To perform …
This notebook paper presents an overview and comparative analysis of our system designed for activity detection in extended videos (ActEV-PC) in ActivityNet Challenge 2019. Specifically, we exploit person/vehicle detections in spatial level and action localization in temporal level for action detection in surveillance videos. The mechanism of different tubelet generation and model decomposition methods are studied as well. The detection results are finally predicted by late fusing the results from each component.
This notebook paper presents an overview and comparative analysis of our systems designed for the following three tasks in ActivityNet Challenge 2019: trimmed action recognition, dense-captioning events in videos, and spatio-temporal action localization.
It is encouraged to see that progress has been made to bridge videos and natural language. However, mainstream video captioning methods suffer from slow inference speed due to the sequential manner of autoregressive decoding, and prefer generating generic descriptions due to the insufficient training of visual words (e.g., nouns and verbs) and inadequate decoding paradigm. In this paper, we propose a non-autoregressive decoding based model with a coarse-to-fine captioning procedure to alleviate these defects. In implementations, we employ a bi-directional self-attention based network as our la…
This paper presents a review of human activity recognition and behaviour understanding in video sequence. The key objective of this paper is to provide a general review on the overall process of a surveillance system used in the current trend. Visual surveillance system is directed on automatic identification of events of interest, especially on tracking and classification of moving objects. The processing step of the video surveillance system includes the following stages: Surrounding model, object representation, object tracking, activity recognition and behaviour understanding. It describes…
This paper proposes a human activity recognition method which is based on features learned from 3D video data without incorporating domain knowledge. The experiments on data collected by RGBD cameras produce results outperforming other techniques. Our feature encoding method follows the bag-of-visual-word model, then we use a SVM classifier to recognise the activities. We do not use skeleton or tracking information and the same technique is applied on color and depth data.
Queen Mary University of London（伦敦玛丽女王大学）；Vision Semantics Limited（视觉语义有限公司）；Adobe Research（Adobe研究院）；Peking University（北京大学）；Beijing Institute for General Artificial Intelligence（北京通用人工智能研究院）
Current methods for video activity localisation over time assume implicitly that activity temporal boundaries labelled for model training are determined and precise. However, in unscripted natural videos, different activities mostly transit smoothly, so that it is intrinsically ambiguous to determine in labelling precisely when an activity starts and ends over time. Such uncertainties in temporal labelling are currently ignored in model training, resulting in learning mis-matched video-text correlation with poor generalisation in test. In this work, we solve this problem by introducing Elastic…
Recently, there has been a growing interest in wearable sensors which provides new research perspectives for 360 ° video analysis. However, the lack of 360 ° datasets in literature hinders the research in this field. To bridge this gap, in this paper we propose a novel Egocentric (first-person) 360° Kinetic human activity video dataset (EgoK360). The EgoK360 dataset contains annotations of human activity with different sub-actions, e.g., activity Ping-Pong with four sub-actions which are pickup-ball, hit, bounce-ball and serve. To the best of our knowledge, EgoK360 is the first dataset in the …
We present the Multiview Extended Video with Activities (MEVA) dataset, a new and very-large-scale dataset for human activity recognition. Existing security datasets either focus on activity counts by aggregating public video disseminated due to its content, which typically excludes same-scene background video, or they achieve persistence by observing public areas and thus cannot control for activity content. Our dataset is over 9300 hours of untrimmed, continuous video, scripted to include diverse, simultaneous activities, along with spontaneous background activity. We have annotated 144 hour…
Domain Generalization for Improved Human Activity Recognition in Office Space Videos Using Adaptive Pre-processing
Bangladesh University of Engineering and Technology (BUET)（孟加拉工程技术大学）；Center for Bioengineering Innovation and Design (CBID)（生物工程创新与设计中心）；Johns Hopkins University（约翰斯·霍普金斯大学）
ActivityForensics: A Comprehensive Benchmark for Localizing Manipulated Activity in Videos
College of Computer Science and Technology, Zhejiang University（浙江大学计算机科学与技术学院）；School of Electrical and Electronic Engineering, Nanyang Technological University（南洋理工大学电气与电子工程学院）；School of Computing and Artificial Intelligence, Jiangxi University of Finance and Economics（江西财经大学计算机与人工智能学院）；Jiangxi Provincial Key Laboratory of Multimedia Intelligent Processing（江西省多媒体智能处理重点实验室）；Faculty of Engineering, Shenzhen MSU-BIT University（深圳北理莫斯科大学工程系）；VinUniversity
The Out the Window (OTW) dataset is a crowdsourced activity dataset containing 5,668 instances of 17 activities from the NIST Activities in Extended Video (ActEV) challenge. These videos are crowdsourced from workers on the Amazon Mechanical Turk using a novel scenario acting strategy, which collects multiple instances of natural activities per scenario. Turkers are instructed to lean their mobile device against an upper story window overlooking an outdoor space, walk outside to perform a scenario involving people, vehicles and objects, and finally upload the video to us for annotation. Perfor…
Human Activity Recognition in RGB-D videos has been an active research topic during the last decade. However, no efforts have been found in the literature, for recognizing human activity in RGB-D videos where several performers are performing simultaneously. In this paper we introduce such a challenging dataset with several performers performing the activities. We present a novel method for recognizing human activities in such videos. The proposed method aims in capturing the motion information of the whole video by producing a dynamic image corresponding to the input video. We use two paralle…
Beijing University of Posts and Telecommunications（北京邮电大学）；Beijing Key Laboratory of Network System and Network Culture（北京网络系统与网络文化重点实验室）
Activity detection in surveillance videos is a challenging task caused by small objects, complex activity categories, its untrimmed nature, etc. Existing methods are generally limited in performance due to inaccurate proposals, poor classifiers or inadequate post-processing method. In this work, we propose a comprehensive and effective activity detection system in untrimmed surveillance videos for person-centered and vehicle-centered activities. It consists of four modules, i.e., object localizer, proposal filter, activity classifier and activity refiner. For person-centered activities, a nove…
Activity recognition in surgical videos is a key research area for developing next-generation devices and workflow monitoring systems. Since surgeries are long processes with highly-variable lengths, deep learning models used for surgical videos often consist of a two-stage setup using a backbone and temporal sequence model. In this paper, we investigate many state-of-the-art backbones and temporal models to find architectures that yield the strongest performance for surgical activity recognition. We first benchmark the models performance on a large-scale activity recognition dataset containin…
Queen Mary University of London（伦敦玛丽女王大学）；WICT, Peking University（北京大学王选计算机研究所）；Adobe Research（Adobe研究院）
Video activity localisation has recently attained increasing attention due to its practical values in automatically localising the most salient visual segments corresponding to their language descriptions (sentences) from untrimmed and unstructured videos. For supervised model training, a temporal annotation of both the start and end time index of each video segment for a sentence (a video moment) must be given. This is not only very expensive but also sensitive to ambiguity and subjective annotation bias, a much harder task than image labelling. In this work, we develop a more accurate weakly…
Video understanding is one of the most challenging topics in computer vision. In this paper, a four-stage video understanding pipeline is presented to simultaneously recognize all atomic actions and the single on-going activity in a video. This pipeline uses objects and motions from the video and a graph-based knowledge representation network as prior reference. Two deep networks are trained to identify objects and motions in each video sequence associated with an action. Low Level image features are then used to identify objects of interest in that video sequence. Confidence scores are assign…
Indiana University（印第安纳大学）；EgoVid Inc.（EgoVid公司）；Jet Propulsion Laboratory, California Institute of Technology（加州理工学院喷气推进实验室）；Xi’an Jiaotong-Liverpool University（西交利物浦大学）；Ulsan National Institute of Science and Technology（蔚山科学技术院）
Privacy protection from surreptitious video recordings is an important societal challenge. We desire a computer vision system (e.g., a robot) that can recognize human activities and assist our daily life, yet ensure that it is not recording video that may invade our privacy. This paper presents a fundamental approach to address such contradicting objectives: human activity recognition while only using extreme low-resolution (e.g., 16x12) anonymized videos. We introduce the paradigm of inverse super resolution (ISR), the concept of learning the optimal set of image transformations to generate m…
We propose an approach for forecasting video of complex human activity involving multiple people. Direct pixel-level prediction is too simple to handle the appearance variability in complex activities. Hence, we develop novel intermediate representations. An architecture combining a hierarchical temporal model for predicting human poses and encoder-decoder convolutional neural networks for rendering target appearances is proposed. Our hierarchical model captures interactions among people by adopting a dynamic group-based interaction mechanism. Next, our appearance rendering network encodes the…
Shenzhen Technology University（深圳技术大学）；Central South University（中南大学）；Nanjing University of Aeronautics and Astronautics（南京航空航天大学）
Automatic smoky vehicle detection in videos is a superior solution to the traditional expensive remote sensing one with ultraviolet-infrared light devices for environmental protection agencies. However, it is challenging to distinguish vehicle smoke from shadow and wet regions coming from rear vehicle or clutter roads, and could be worse due to limited annotated data. In this paper, we first introduce a real-world large-scale smoky vehicle dataset with 75,000 annotated smoky vehicle images, facilitating the effective training of advanced deep learning models. To enable fair algorithm compariso…
Beijing University of Technology（北京工业大学）；The University of Sydney（悉尼大学）；Dalian University of Technology（大连理工大学）
In multi-camera video surveillance, it is challenging to represent videos from different cameras properly and fuse them efficiently for specific applications such as human activity recognition and clustering. In this paper, a novel representation for multi-camera video data, namely the Product Grassmann Manifold (PGM), is proposed to model video sequences as points on the Grassmann manifold and integrate them as a whole in the product manifold form. Additionally, with a new geometry metric on the product manifold, the conventional Low Rank Representation (LRR) model is extended onto PGM and th…
Bangladesh University of Engineering and Technology (BUET)（孟加拉国工程技术大学）
In the advent of wearable body-cameras, human activity classification from First-Person Videos (FPV) has become a topic of increasing importance for various applications, including in life-logging, law-enforcement, sports, workplace, and healthcare. One of the challenging aspects of FPV is its exposure to potentially sensitive objects within the user's field of view. In this work, we developed a privacy-aware activity classification system focusing on office videos. We utilized a Mask-RCNN with an Inception-ResNet hybrid as a feature extractor for detecting, and then blurring out sensitive obj…
Human activity recognition is an emerging and important area in computer vision which seeks to determine the activity an individual or group of individuals are performing. The applications of this field ranges from generating highlight videos in sports, to intelligent surveillance and gesture recognition. Most activity recognition systems rely on a combination of convolutional neural networks (CNNs) to perform feature extraction from the data and recurrent neural networks (RNNs) to determine the time dependent nature of the data. This paper proposes and designs two transformer neural networks …
University of Science and Technology of China（中国科学技术大学）；Southeast University（东南大学）；Alibaba Group（阿里巴巴集团）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；CAS Center for Excellence in Brain Science and Intelligence Technology（中国科学院脑科学与智能技术卓越创新中心）
Current metrics for video captioning are mostly based on the text-level comparison between reference and candidate captions. However, they have some insuperable drawbacks, e.g., they cannot handle videos without references, and they may result in biased evaluation due to the one-to-many nature of video-to-text and the neglect of visual relevance. From the human evaluator's viewpoint, a high-quality caption should be consistent with the provided video, but not necessarily be similar to the reference in literal or semantics. Inspired by human evaluation, we propose EMScore (Embedding Matching-ba…
Current approaches for activity recognition often ignore constraints on computational resources: 1) they rely on extensive feature computation to obtain rich descriptors on all frames, and 2) they assume batch-mode access to the entire test video at once. We propose a new active approach to activity recognition that prioritizes "what to compute when" in order to make timely predictions. The main idea is to learn a policy that dynamically schedules the sequence of features to compute on selected frames of a given test video. In contrast to traditional static feature selection, our approach cont…
University of Surrey（萨里大学）；National Institute of Technology（国家理工学院）
Video activity recognition by deep neural networks is impressive for many classes. However, it falls short of human performance, especially for challenging to discriminate activities. Humans differentiate these complex activities by recognising critical spatio-temporal relations among explicitly recognised objects and parts, for example, an object entering the aperture of a container. Deep neural networks can struggle to learn such critical relationships effectively. Therefore we propose a more human-like approach to activity recognition, which interprets a video in sequential temporal phases …
In this paper, we newly introduce the concept of temporal attention filters, and describe how they can be used for human activity recognition from videos. Many high-level activities are often composed of multiple temporal parts (e.g., sub-events) with different duration/speed, and our objective is to make the model explicitly learn such temporal structure using multiple attention filters and benefit from them. Our temporal filters are designed to be fully differentiable, allowing end-of-end training of the temporal filters together with the underlying frame-based or segment-based convolutional…
Rutgers University–New Brunswick（罗格斯大学新布朗斯维克分校）；Amazon Web Service（亚马逊网络服务）
Multi-label activity recognition is designed for recognizing multiple activities that are performed simultaneously or sequentially in each video. Most recent activity recognition networks focus on single-activities, that assume only one activity in each video. These networks extract shared features for all the activities, which are not designed for multi-label activities. We introduce an approach to multi-label activity recognition that extracts independent feature descriptors for each activity and learns activity correlations. This structure can be trained end-to-end and plugged into any exis…
Human activities comprise several sub-activities performed in a sequence and involve interactions with various objects. This makes reasoning about the object affordances a central task for activity recognition. In this work, we consider the problem of jointly labeling the object affordances and human activities from RGB-D videos. We frame the problem as a Markov Random Field where the nodes represent objects and sub-activities, and the edges represent the relationships between object affordances, their relations with sub-activities, and their evolution over time. We formulate the learning prob…
Group activity detection in soccer can be done by using either video data or player and ball trajectory data. In current soccer activity datasets, activities are labelled as atomic events without a duration. Given that the state-of-the-art activity detection methods are not well-defined for atomic actions, these methods cannot be used. In this work, we evaluated the effectiveness of activity recognition models for detecting such events, by using an intuitive non-maximum suppression process and evaluation metrics. We also considered the problem of explicitly modeling interactions between player…
This note describes the details of our solution to the dense-captioning events in videos task of ActivityNet Challenge 2018. Specifically, we solve this problem with a two-stage way, i.e., first temporal event proposal and then sentence generation. For temporal event proposal, we directly leverage the three-stage workflow in [13, 16]. For sentence generation, we capitalize on LSTM-based captioning framework with temporal attention mechanism (dubbed as LSTM-T). Moreover, the input visual sequence to the LSTM-based video captioning model is comprised of RGB and optical flow images. At inference,…
University of Chinese Academy of Sciences（中国科学院大学）；Institute of Deep Learning, Baidu Research（百度研究院深度学习研究所）；National Engineering Laboratory for Deep Learning Technology and Application（深度学习技术及应用国家工程实验室）
Facial expression recognition plays an important role in human-computer interaction. In this paper, we propose the Coarse-to-Fine Cascaded network with Smooth Predicting (CFC-SP) to improve the performance of facial expression recognition. CFC-SP contains two core components, namely Coarse-to-Fine Cascaded networks (CFC) and Smooth Predicting (SP). For CFC, it first groups several similar emotions to form a rough category, and then employs a network to conduct a coarse but accurate classification. Later, an additional network for these grouped emotions is further used to obtain fine-grained pr…
Rochester Institute of Technology（罗切斯特理工学院）；Kodak Alaris（柯达伊士曼）
In this study, we develop an unsupervised coarse-to-fine video analysis framework and prototype system to extract a salient object in a video sequence. This framework starts from tracking grid-sampled points along temporal frames, typically using KLT tracking method. The tracking points could be divided into several groups due to their inconsistent movements. At the same time, the SLIC algorithm is extended into 3D space to generate supervoxels. Coarse segmentation is achieved by combining the categorized tracking points and supervoxels of the corresponding frame in the video sequence. Finally…
Creating and labelling datasets of videos for use in training Human Activity Recognition models is an arduous task. In this paper, we approach this by using 3D rendering tools to generate a synthetic dataset of videos, and show that a classifier trained on these videos can generalise to real videos. We use five different augmentation techniques to generate the videos, leading to a wide variety of accurately labelled unique videos. We fine tune a pre-trained I3D model on our videos, and find that the model is able to achieve a high accuracy of 73% on the HMDB51 dataset over three classes. We al…
This thesis explore different approaches using Convolutional and Recurrent Neural Networks to classify and temporally localize activities on videos, furthermore an implementation to achieve it has been proposed. As the first step, features have been extracted from video frames using an state of the art 3D Convolutional Neural Network. This features are fed in a recurrent neural network that solves the activity classification and temporally location tasks in a simple and flexible way. Different architectures and configurations have been tested in order to achieve the best performance and learni…
Human video motion transfer (HVMT) aims to synthesize videos that one person imitates other persons' actions. Although existing GAN-based HVMT methods have achieved great success, they either fail to preserve appearance details due to the loss of spatial consistency between synthesized and exemplary images, or generate incoherent video results due to the lack of temporal consistency among video frames. In this paper, we propose Coarse-to-Fine Flow Warping Network (C2F-FWN) for spatial-temporal consistent HVMT. Particularly, C2F-FWN utilizes coarse-to-fine flow warping and Layout-Constrained De…
University College London（伦敦大学学院）；Dithen（Dithen公司）
We investigate video classification via a two-stream convolutional neural network (CNN) design that directly ingests information extracted from compressed video bitstreams. Our approach begins with the observation that all modern video codecs divide the input frames into macroblocks (MBs). We demonstrate that selective access to MB motion vector (MV) information within compressed video bitstreams can also provide for selective, motion-adaptive, MB pixel decoding (a.k.a., MB texture decoding). This in turn allows for the derivation of spatio-temporal video activity regions at extremely high spe…
