scholarly journals Extending Appearance Based Gait Recognition with Depth Data

2019 ◽  
Vol 9 (24) ◽  
pp. 5529
Author(s):  
Kristijan Lenac ◽  
Diego Sušanj ◽  
Adnan Ramakić ◽  
Domagoj Pinčić
Keyword(s):  
Information Fusion ◽  
Recognition Accuracy ◽  
Gait Recognition ◽  
Computational Cost ◽  
Video Stream ◽  
Depth Data ◽  
Video Images ◽  
Extraction Step ◽  
Different Types ◽  
Subject Identification

Each individual describes unique patterns during their gait cycles. This information can be extracted from the live video stream and used for subject identification. In appearance based recognition methods, this is done by tracking silhouettes of persons across gait cycles. In recent years, there has been a profusion of sensors that in addition to RGB video images also provide depth data in real-time. When such sensors are used for gait recognition, existing RGB appearance based methods can be extended to get a substantial gain in recognition accuracy. In this paper, this is accomplished using information fusion techniques that combine features from extracted silhouettes, used in traditional appearance based methods, and the height feature that can now be estimated using depth data. The latter is estimated during the silhouette extraction step with minimal additional computational cost. Two approaches are proposed that can be implemented easily as an extension to existing appearance based methods. An extensive experimental evaluation was performed to provide insights into how much the recognition accuracy can be improved. The results are presented and discussed considering different types of subjects and populations of different height distributions.

Depth-Based Real-Time Gait Recognition

2020 ◽  
Vol 29 (16) ◽  
pp. 2050266
Author(s):  
Adnan Ramakić ◽  
Diego Sušanj ◽  
Kristijan Lenac ◽  
Zlatko Bundalo
Keyword(s):  
Real Time ◽  
Gait Recognition ◽  
Video Stream ◽  
Outdoor Environment ◽  
Microsoft Kinect ◽  
Recognition Method ◽  
Real Time Processing ◽  
Depth Data ◽  
Video Images ◽  
Subject Identification

Each person describes unique patterns during gait cycles and this information can be extracted from live video stream and used for subject identification. In recent years, there has been a profusion of sensors that in addition to RGB video images also provide depth data in real-time. In this paper, a method to enhance the appearance-based gait recognition method by also integrating features extracted from depth data is proposed. Two approaches are proposed that integrate simple depth features in a way suitable for real-time processing. Unlike previously presented works which usually use a short range sensors like Microsoft Kinect, here, a long-range stereo camera in outdoor environment is used. The experimental results for the proposed approaches show that recognition rates are improved when compared to existing popular gait recognition methods.

Gait Period Detection Based on Layered Coding of Depth Information

2014 ◽  
Vol 644-650 ◽  
pp. 1015-1018 ◽  
Author(s):  
Hao Lin Zhang ◽  
Xian Ye Ben ◽  
Peng Zhang ◽  
Tian Jiao Liu
Keyword(s):  
Detection Method ◽  
Recognition Accuracy ◽  
Gait Recognition ◽  
Video Stream ◽  
Depth Information ◽  
Layered Coding ◽  
Depth Images ◽  
Gait Features ◽  
Period Detection ◽  
High Recognition Accuracy

Gait period detection, serving as a preprocessor for gait recognition, is commonly studied in the recent past. In this paper, we proposed a novel gait period detection method for depth gait video stream. The method introduces the concept of layered coding for depth images which decreases computational complexity. Furthermore, the extreme value of the sum of layered codes for gait sequence is utilized to judge the period endpoint, which is in accord with the naked-eye observation. In addition, gait recognition experiments on the TUM GAID database are conducted with the description of gait features of one single detected period by the proposed scheme using tensor representation. The high recognition accuracy verifies the effectiveness of the proposed depth gait period detection method.

scholarly journals Methods for Preventing Visual Attacks in Convolutional Neural Networks Based on Data Discard and Dimensionality Reduction

2021 ◽  
Vol 11 (11) ◽  
pp. 5235
Author(s):  
Nikita Andriyanov
Keyword(s):  
Neural Network ◽  
Convolutional Neural Network ◽  
Network Inference ◽  
Recognition Accuracy ◽  
Network Architectures ◽  
Reduced Dimensions ◽  
Different Types ◽  
Rectangular Area ◽  
Noise Impulse ◽  
Selection Of

The article is devoted to the study of convolutional neural network inference in the task of image processing under the influence of visual attacks. Attacks of four different types were considered: simple, involving the addition of white Gaussian noise, impulse action on one pixel of an image, and attacks that change brightness values within a rectangular area. MNIST and Kaggle dogs vs. cats datasets were chosen. Recognition characteristics were obtained for the accuracy, depending on the number of images subjected to attacks and the types of attacks used in the training. The study was based on well-known convolutional neural network architectures used in pattern recognition tasks, such as VGG-16 and Inception_v3. The dependencies of the recognition accuracy on the parameters of visual attacks were obtained. Original methods were proposed to prevent visual attacks. Such methods are based on the selection of “incomprehensible” classes for the recognizer, and their subsequent correction based on neural network inference with reduced image sizes. As a result of applying these methods, gains in the accuracy metric by a factor of 1.3 were obtained after iteration by discarding incomprehensible images, and reducing the amount of uncertainty by 4–5% after iteration by applying the integration of the results of image analyses in reduced dimensions.

scholarly journals Technological video survaillance as protective tool of smart electricity meters from unauthorized access

2020 ◽  
Vol 2020 (2) ◽  
pp. 7-13
Author(s):  
Oxana Garifullaevna Koispaeva ◽  
Sergey Vladimirovich Golovko ◽  
Maksim Almansurovich Nadeev
Keyword(s):  
Video Surveillance ◽  
Surveillance System ◽  
Information Technologies ◽  
Video Recording ◽  
Video Stream ◽  
Recording System ◽  
Economic Feasibility ◽  
Electrical Networks ◽  
New Information ◽  
Different Types

The article touches upon the problems of interference in the electricity metering system and different types of electricity loss underestimation. It has been stated that in modern conditions the only effective way to identify and prevent violations is to install a remote electricity meter at the borderline of balance and operational responsibilities. РиМ 384.02 smart electricity meter is considered as an example. A picture of the device is presented and its main positive characteristics are listed. It has been proposed to introduce the twenty-four-hour video surveillance and video recording system into an expensive smart electricity meter in order to maintain the integrity and operability of the device. Video stream processing by means of machine vision and machine learning has been proposed. The economic feasibility of digitalization of the technological video surveillance system is being substantiated. The need of optimization of the electrical networks, the improvement of electricity metering system and of introducing the new information technologies in energy sales has been substantiated

scholarly journals Body Joints and Trajectory Guided 3D Deep Convolutional Descriptors for Human Activity Identification

2019 ◽  
Vol 8 (12) ◽  
pp. 1016-1021
Keyword(s):  
Optical Flow ◽  
Human Activity ◽  
Large Scale ◽  
Recognition Accuracy ◽  
Computational Cost ◽  
The Body ◽  
Support Vector ◽  
3D Network ◽  
Body Joints ◽  
Body Joint

Human Activity Identification (HAI) in videos is one of the trendiest research fields in the computer visualization. Among various HAI techniques, Joints-pooled 3D-Deep convolutional Descriptors (JDD) have achieved effective performance by learning the body joint and capturing the spatiotemporal characteristics concurrently. However, the time consumption for estimating the locale of body joints by using large-scale dataset and computational cost of skeleton estimation algorithm were high. The recognition accuracy using traditional approaches need to be improved by considering both body joints and trajectory points together. Therefore, the key goal of this work is to improve the recognition accuracy using an optical flow integrated with a two-stream bilinear model, namely Joints and Trajectory-pooled 3D-Deep convolutional Descriptors (JTDD). In this model, an optical flow/trajectory point between video frames is also extracted at the body joint positions as input to the proposed JTDD. For this reason, two-streams of Convolutional 3D network (C3D) multiplied with the bilinear product is used for extracting the features, generating the joint descriptors for video sequences and capturing the spatiotemporal features. Then, the whole network is trained end-to-end based on the two-stream bilinear C3D model to obtain the video descriptors. Further, these video descriptors are classified by linear Support Vector Machine (SVM) to recognize human activities. Based on both body joints and trajectory points, action recognition is achieved efficiently. Finally, the recognition accuracy of the JTDD model and JDD model are compared.

Research on the Strategy of Multi Sensor Information Fusion in Coal Mine Disaster Scene

2012 ◽  
Vol 220-223 ◽  
pp. 1929-1933
Author(s):  
Yu Qing Wang ◽  
Yu Xin Qin ◽  
Zhi Guo Li ◽  
Le Deng
Keyword(s):  
Information Fusion ◽  
Coal Mine ◽  
Coal Mines ◽  
Outburst Flood ◽  
Roof Collapse ◽  
Gas Outburst ◽  
Different Types ◽  
Sensor Information

Different types of disasters occur frequently in coal mines. This paper analyzed the characteristics of different disasters, chosen the corresponding sensors to collect the information of disaster scene, and discussed the methods of multi sensor information fusion. Lastly, the multi-sensor information fusion strategies for fire, gas outburst, flood, and roof collapse were proposed in this research.

scholarly journals Elderly Fall Detection with an Accelerometer Using Lightweight Neural Networks

Electronics ◽  
2019 ◽  
Vol 8 (11) ◽  
pp. 1354 ◽  
Author(s):  
Gaojing Wang ◽  
Qingquan Li ◽  
Lei Wang ◽  
Yuanshi Zhang ◽  
Zheng Liu
Keyword(s):  
Neural Networks ◽  
Computational Cost ◽  
Fall Detection ◽  
The Elderly ◽  
Research Problem ◽  
Wearable Devices ◽  
The Past ◽  
Machine Learning Methods ◽  
Open Research ◽  
Different Types

Falls have been one of the main threats to people’s health, especially for the elderly. Detecting falls in time can prevent the long lying time, which is extremely fatal. This paper intends to show the efficacy of detecting falls using a wearable accelerometer. In the past decade, the fall detection problem has been extensively studied. However, since the hardware resources of wearable devices are limited, designing highly accurate embeddable models with feasible computational cost remains an open research problem. In this paper, different types of shallow and lightweight neural networks, including supervised and unsupervised models are explored to improve the fall detection results. Experiment results on a large open dataset show that the lightweight neural networks proposed have obtained much better results than machine learning methods used in previous work. Moreover, the storage and computation requirements of these lightweight models are only a few hundredths of deep neural networks in literature. In tested lightweight neural networks, the best one is proved to be the supervised convolutional neural network (CNN) that can achieve an accuracy beyond 99.9% with only 441 parameters. Its storage and computation requirements are only 1.2 KB and 0.008 MFLOPs, which make it more suitable to be implemented in wearable devices with restricted memory size and computation power.

scholarly journals Distributed Radar Information Fusion for Gait Recognition and Fall Detection

2020 ◽  
Author(s):  
Haobo Li ◽  
Julien le Kernec ◽  
Ajay Mehul ◽  
Sevgi Zubeyde Gurbuz ◽  
Francesco Fioranelli
Keyword(s):  
Information Fusion ◽  
Gait Recognition ◽  
Fall Detection ◽  
Radar Information

The Design of Computer Forms for Tropical Medical Research

1986 ◽  
Vol 25 (04) ◽  
pp. 229-232 ◽  
Author(s):  
P. Byass
Keyword(s):  
Medical Research ◽  
Relational Database ◽  
Research Data ◽  
Computer Hardware ◽  
Management Systems ◽  
Different Types ◽  
Subject Identification ◽  
The Tropics ◽  
Relational Database Management ◽  
Relational Database Management Systems

SummaryThe design of forms for recording medical research data in the field is discussed in the light of the computer hardware and software to be used for their subsequent analysis. In particular, the principles of relational database management systems are extrapolated to the structure of forms. Specific issues of subject identification, layout of forms, different types of data and coding methods are developed. Finally, the special problems of applying these principles in the tropics, and experiences of so doing, are considered.

scholarly journals IoT-IE: An Information-Entropy-Based Approach to Traffic Anomaly Detection in Internet of Things

2021 ◽  
Vol 2021 ◽  
pp. 1-13
Author(s):  
Yizhen Sun ◽  
Jianjiang Yu ◽  
Jianwei Tian ◽  
Zhongwei Chen ◽  
Weiping Wang ◽  
...  
Keyword(s):  
Internet Of Things ◽  
Information Entropy ◽  
Recognition Accuracy ◽  
False Positive Rate ◽  
Security Issues ◽  
Different Types ◽  
Positive Rate ◽  
Iot Devices ◽  
Abnormal Behaviors ◽  
Traffic Anomaly

Security issues related to the Internet of Things (IoTs) have attracted much attention in many fields in recent years. One important problem in IoT security is to recognize the type of IoT devices, according to which different strategies can be designed to enhance the security of IoT applications. However, existing IoT device recognition approaches rarely consider traffic attacks, which might change the pattern of traffic and consequently decrease the recognition accuracy of different IoT devices. In this work, we first validate by experiments that traffic attacks indeed decrease the recognition accuracy of existing IoT device recognition approaches; then, we propose an approach called IoT-IE that combines information entropy of different traffic features to detect traffic anomaly. We then enhance the robustness of IoT device recognition by detecting and ignoring the abnormal traffic detected by our approach. Experimental evaluations show that IoT-IE can effectively detect abnormal behaviors of IoT devices in the traffic under eight different types of attacks, achieving a high accuracy value of 0.977 and a low false positive rate of 0.011. It also achieves an accuracy of 0.969 in a multiclassification experiment with 7 different types of attacks.

Sign in / Sign up

Export Citation Format

Share Document