Search

The increasing availability of wireless access points (APs) is leading toward human sensing applications based on Wi-Fi signals as support or alternative tools to the widespread visual sensors, where the signals enable to address well-known vision-related problems such as illumination changes or occlusions. Indeed, using image synthesis techniques to translate radio frequencies to the visible spectrum can become essential to obtain otherwise unavailable visual data. This domain-to-domain translation is feasible because both objects and people affect electromagnetic waves, causing radio and optical frequencies variations. In the literature, models capable of inferring radio-to-visual features mappings have gained momentum in the last few years since frequency changes can be observed in the radio domain through the channel state information (CSI) of Wi-Fi APs, enabling signal-based feature extraction, e.g. amplitude. On this account, this paper presents a novel two-branch generative neural network that effectively maps radio data into visual features, following a teacher–student design that exploits a cross-modality supervision strategy. The latter conditions signal-based features in the visual domain to completely replace visual data. Once trained, the proposed method synthesizes human silhouette and skeleton videos using exclusively Wi-Fi signals. The approach is evaluated on publicly available data, where it obtains remarkable results for both silhouette and skeleton videos generation, demonstrating the effectiveness of the proposed cross-modality supervision strategy. [ABSTRACT FROM AUTHOR]

Falls are reported to be the leading causes of accidental deaths among elderly people. Automatic detection of falls from video sequences is an assistant technology for low-cost health care systems. In this paper, we present a novel slow feature analysis based framework for fall detection in a house care environment. Firstly, a foreground human body is extracted by a background subtraction technique. After morphological operations, the human silhouette is refined and covered by a fitted ellipse. Secondly, six shape features are quantified from the covered silhouette to represent different human postures. With the help of the learned slow feature functions, the shape feature sequences are transformed into slow feature sequences with discriminative information about human actions. To represent the fall incidents, the squared first order temporal derivatives of the slow features are accumulated into a classification vector. Lastly, falls are distinguished from other daily actions, such as walking, crouching, and sitting, by the trained directed acyclic graph support vector machine. Experiments on the multiple-camera fall dataset and the SDUFall dataset demonstrate that our method is comparable to other state-of-the-art methods, achieving 94.00% recognition rate on the former dataset and 96.57% on the latter one. [ABSTRACT FROM AUTHOR]

Highlights • An effective background subtraction technique is proposed. • A novel view-independent CNNs classifier which is applied. • High-quality network inputs have low computational cost. • A simple voting classifier works fairly well in multi-camera system. Abstract In this paper, a computer vision based framework is proposed that detects falls from surveillance videos. Firstly, we employ background subtraction and rank pooling to model spatial and temporal representations in videos, respectively. We then introduce a novel three-stream Convolutional Neural Networks as an event classifier. Silhouettes and their motion history images serve as input to the first two streams, while dynamic images whose temporal duration is equal to motion history images, are used as input to the third stream. Finally, we apply voting on the results of event classification to perform multi-camera fall detection. The main novelty of our method against the conventional ones is that high-quality spatiotemporal representations in different levels are learned to take full advantage of the appearance and motion information. Extensive experiments have been conducted on two widely used fall datasets. The results have shown to demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Many countries are experiencing a rapid increase in their elderly populations, increasing the demand for appropriate healthcare systems including fall-detection systems. In recent years, many fall-detection systems have been developed, although most require the use of wearable devices. Such systems function only when the subject is wearing the device. A vision-based system presents a more convenient option. However, visual features typically depend on camera view; a single, fixed camera may not properly identify falls occurring in various directions. Thus, this study presents a solution that involves using multiple cameras. The study offers two main contributions. First, in contrast to most vision-based systems that analyze silhouettes to detect falls, the present system proposes a novel feature for measuring the degree of impact shock that is easily detectable with a wearable device but more difficult with a computer vision system. In addition, the degree of impact shock is less sensitive to camera views and can be extracted more robustly than a silhouette. Second, the proposed method uses a majority-voting strategy based on multiple views to avoid performing the tedious camera calibration required by most multiple-camera approaches. Specifically, the proposed method is based on spatio-temporal interest points (STIPs). The number of local STIP clusters is designed to indicate the degree of impact shock and body vibration. Sequences of these features are concatenated into feature vectors that are then fed into a support vector machine to classify the fall event. A majority-voting strategy based on multiple views is then used for the final determination. The proposed method has been applied to a publicly available dataset to offer evidence that the proposed method outperforms existing methods based on the same data input. [ABSTRACT FROM AUTHOR]

Robust tracking of a human in a video sequence is an essential prerequisite to an increasing number of applications, where a robot needs to interact with a human user or operates in a human-inhabited environment. This paper presents a robust approach that enables a mobile robot to detect and track a human using an onboard RGB-D sensor. Such robots could be used for security, surveillance, and assistive robotics applications. The proposed approach has real-time computation power through a unique combination of new ideas and well-established techniques. In the proposed method, background subtraction is combined with depth segmentation detector and template matching method to initialize the human tracking automatically. A novel concept of head and hand creation based on depth of interest is introduced in this paper to track the human silhouette in a dynamic environment, when the robot is moving. To make the algorithm robust, a series of detectors (e.g., height, size, and shape) is utilized to distinguish target human from other objects. Because of the relatively high computation time of the silhouette-matching-based method, a confidence level is defined, which allows using the matching-based method only where it is imperative. An unscented Kalman filter is used to predict the human location in the image frame to maintain the continuity of the robot motion. The efficacy of the approach is demonstrated through a real experiment on a mobile robot navigating in an indoor environment. [ABSTRACT FROM PUBLISHER]

Human silhouette extraction has been a primary step to estimate human poses or classify activities from videos. While the accuracy of human silhouettes has great impact on the follow-on human pose/gait estimation, it has been important to guarantee the highly-accurate extraction of human silhouettes. However, traditional methods such as motion segmentation can be fragile due to the complexity of real-world environment. In this paper, we propose an automated human silhouette extraction algorithm to attain this highly-demanded task. In our proposed scheme, the initial motion segmentation of foreground objects was roughly computed by Stauffer’s background subtraction using Gaussian mixtures, and then refined by the proposed Laplacian fitting scheme. In our method, the candidate regions of human objects are taken as the initial input, their Laplacian matrices are constructed, and Eigen mattes are then obtained by minimizing on Laplacian matrices. RANSAC algorithm is then applied to fit the Eigen mattes iteratively with inliers of the initially estimated motion blob. Finally, the foreground human silhouettes are obtained from the optimized matte fitting. Experimental results on a number of test videos validated that the proposed Laplacian fitting scheme enhances the accuracy in automated human silhouette extraction, exhibiting a potential use of our Laplacian fitting algorithm in many silhouette-based applications such as human pose estimation. [ABSTRACT FROM AUTHOR]

Recognizing actions from a monocular video is a very hot topic in computer vision recently. In this paper, we propose a new representation of actions, the co-occurrence matrices descriptor, on the intrinsic shape manifold learned by graph embedding. The co-occurrence matrices descriptor captures more temporal information than the bag of words (histogram) descriptor which only considers the spatial information, thus boosting the classification accuracy. In addition, we compare the performance of the co-occurrence matrices descriptor on different manifolds learned by various graph-embedding methods. Graph-embedding methods preserve as much of the significant structure of the high-dimensional data as possible in the low-dimensional map. The results show that nonlinear algorithms are more robust than linear ones. Furthermore, we conclude that the label information plays a critical role in learning more discriminating manifolds. [ABSTRACT FROM AUTHOR]

Gender is an important cue in social activities. In this correspondence, we present a study and analysis of gender classification based on human gait. Psychological experiments were carried out. These experiments showed that humans can recognize gender based on gait information, and that contributions of different body components vary. The prior knowledge extracted from the psychological experiments can be combined with an automatic method to further improve classification accuracy. The proposed method which combines human knowledge achieves higher performance than some other methods, and is even more accurate than human observers. We also present a numerical analysis of the contributions of different human components, which shows that head and hair, back, chest and thigh are more discriminative than other components. We also did challenging cross-race experiments that used Asian gait data to classify the gender of Europeans, and vice versa. Encouraging results were obtained. All the above prove that gait-based gender classification is feasible in controlled environments. In real applications, it still suffers from many difficulties, such as view variation, clothing and shoes changes, or carrying objects. We analyze the difficulties and suggest some possible solutions. [ABSTRACT FROM AUTHOR]

This paper proposes the use of the FASSD-Net model for semantic segmentation of human silhouettes, these silhouettes can later be used in various applications that require specific characteristics of human interaction observed in video sequences for the understanding of human activities or for human identification. These applications are classified as high-level task semantic understanding. Since semantic segmentation is presented as one solution for human silhouette extraction, it is concluded that convolutional neural networks (CNN) have a clear advantage over traditional methods for computer vision, based on their ability to learn the representations of appropriate characteristics for the task of segmentation. In this work, the FASSD-Net model is used as a novel proposal that promises real-time segmentation in high-resolution images exceeding 20 FPS. To evaluate the proposed scheme, we use the Cityscapes database, which consists of sundry scenarios that represent human interaction with its environment (these scenarios show the semantic segmentation of people, difficult to solve, that favors the evaluation of our proposal), To adapt the FASSD-Net model to human silhouette semantic segmentation, the indexes of the 19 classes traditionally proposed for Cityscapes were modified, leaving only two labels: One for the class of interest labeled as person and one for the background. The Cityscapes database includes the category "human" composed for "rider" and "person" classes, in which the rider class contains incomplete human silhouettes due to self-occlusions for the activity or transport used. For this reason, we only train the model using the person class rather than human category. The implementation of the FASSD-Net model with only two classes shows promising results in both a qualitative and quantitative manner for the segmentation of human silhouettes. [ABSTRACT FROM AUTHOR]