scholarly journals Application of Crowd Simulations in the Evaluation of Tracking Algorithms

Sensors ◽  
2020 ◽  
Vol 20 (17) ◽  
pp. 4960
Author(s):  
Michał Staniszewski ◽  
Paweł Foszner ◽  
Karol Kostorz ◽  
Agnieszka Michalczuk ◽  
Kamil Wereszczyński ◽  
...  
Keyword(s):  
Action Recognition ◽  
Computer Game ◽  
Weather Conditions ◽  
Video Data ◽  
Manual Annotation ◽  
Human Errors ◽  
System A ◽  
Recognition Algorithms ◽  
Crowd Simulations ◽  
Urban Activities

Tracking and action-recognition algorithms are currently widely used in video surveillance, monitoring urban activities and in many other areas. Their development highly relies on benchmarking scenarios, which enable reliable evaluations/improvements of their efficiencies. Presently, benchmarking methods for tracking and action-recognition algorithms rely on manual annotation of video databases, prone to human errors, limited in size and time-consuming. Here, using gained experiences, an alternative benchmarking solution is presented, which employs methods and tools obtained from the computer-game domain to create simulated video data with automatic annotations. Presented approach highly outperforms existing solutions in the size of the data and variety of annotations possible to create. With proposed system, a potential user can generate a sequence of random images involving different times of day, weather conditions, and scenes for use in tracking evaluation. In the design of the proposed tool, the concept of crowd simulation is used and developed. The system is validated by comparisons to existing methods.

scholarly journals Performance of an Advanced Intelligent Control Strategy in a Dynamic Positioning (DP) System Applied to a Semisubmersible Drilling Platform

2021 ◽  
Vol 9 (4) ◽  
pp. 399
Author(s):  
Mohamad Alremeihi ◽  
Rosemary Norman ◽  
Kayvan Pazouki ◽  
Arun Dev ◽  
Musa Bashir
Keyword(s):  
Oil And Gas ◽  
Weather Conditions ◽  
Dynamic Positioning ◽  
System Failures ◽  
Oil Drilling ◽  
System A ◽  
Advanced Control ◽  
Safety And Reliability ◽  
Drilling Operations ◽  
Shallow Seas

Oil drilling and extraction platforms are currently being used in many offshore areas around the world. Whilst those operating in shallow seas are secured to the seabed, for deeper water operations, Dynamic Positioning (DP) is essential for the platforms to maintain their position within a safe zone. Operating DP requires intelligent and reliable control systems. Nearly all DP accidents have been caused by a combination of technical and human failures; however, according to the International Marine Contractors Association (IMCA) DP Incidents Analysis, DP control and thruster system failures have been the leading causes of incidents over the last ten years. This paper will investigate potential operational improvements for DP system accuracy by adding a Predictive Neural Network (PNN) control algorithm in the thruster allocation along with a nonlinear Proportional Integral derivative (PID) motion control system. A DP system’s performance on a drilling platform in oil and gas deep-water fields and subject to real weather conditions is simulated with these advanced control methods. The techniques are developed for enhancing the safety and reliability of DP operations to improve the positioning accuracy, which may allow faster response to a critical situation during DP drilling operations. The semisubmersible drilling platform’s simulation results using the PNN strategy show improved control of the platform’s positioning.

scholarly journals Low-Cost Embedded System Using Convolutional Neural Networks-Based Spatiotemporal Feature Map for Real-Time Human Action Recognition

2021 ◽  
Vol 11 (11) ◽  
pp. 4940
Author(s):  
Jinsoo Kim ◽  
Jeongho Cho
Keyword(s):  
Embedded System ◽  
Real Time ◽  
Action Recognition ◽  
Processing Speed ◽  
Recognition Accuracy ◽  
Low Cost ◽  
Human Action Recognition ◽  
Human Action ◽  
Video Data ◽  
Feature Maps

The field of research related to video data has difficulty in extracting not only spatial but also temporal features and human action recognition (HAR) is a representative field of research that applies convolutional neural network (CNN) to video data. The performance for action recognition has improved, but owing to the complexity of the model, some still limitations to operation in real-time persist. Therefore, a lightweight CNN-based single-stream HAR model that can operate in real-time is proposed. The proposed model extracts spatial feature maps by applying CNN to the images that develop the video and uses the frame change rate of sequential images as time information. Spatial feature maps are weighted-averaged by frame change, transformed into spatiotemporal features, and input into multilayer perceptrons, which have a relatively lower complexity than other HAR models; thus, our method has high utility in a single embedded system connected to CCTV. The results of evaluating action recognition accuracy and data processing speed through challenging action recognition benchmark UCF-101 showed higher action recognition accuracy than the HAR model using long short-term memory with a small amount of video frames and confirmed the real-time operational possibility through fast data processing speed. In addition, the performance of the proposed weighted mean-based HAR model was verified by testing it in Jetson NANO to confirm the possibility of using it in low-cost GPU-based embedded systems.

scholarly journals Cloud Monitoring for Solar Plants with Support Vector Machine Based Fault Detection System

2014 ◽  
Vol 2014 ◽  
pp. 1-10 ◽  
Author(s):  
Hong-Chan Chang ◽  
Shang-Chih Lin ◽  
Cheng-Chien Kuo ◽  
Hao-Ping Yu
Keyword(s):  
Support Vector Machine ◽  
Information System ◽  
Monitoring System ◽  
Solar Power ◽  
Video Data ◽  
Failure Diagnosis ◽  
Support Vector ◽  
Management Platform ◽  
System A ◽  
Cloud Monitoring

This study endeavors to develop a cloud monitoring system for solar plants. This system incorporates numerous subsystems, such as a geographic information system, an instantaneous power-consumption information system, a reporting system, and a failure diagnosis system. Visual C# was integrated with ASP.NET and SQL technologies for the proposed monitoring system. A user interface for database management system was developed to enable users to access solar power information and management systems. In addition, by using peer-to-peer (P2P) streaming technology and audio/video encoding/decoding technology, real-time video data can be transmitted to the client end, providing instantaneous and direct information. Regarding smart failure diagnosis, the proposed system employs the support vector machine (SVM) theory to train failure mathematical models. The solar power data are provided to the SVM for analysis in order to determine the failure types and subsequently eliminate failures at an early stage. The cloud energy-management platform developed in this study not only enhances the management and maintenance efficiency of solar power plants but also increases the market competitiveness of solar power generation and renewable energy.

scholarly journals Recognition of Common Non-Normal Walking Actions Based on Relief-F Feature Selection and Relief-Bagging-SVM

Sensors ◽  
2020 ◽  
Vol 20 (5) ◽  
pp. 1447
Author(s):  
Pan Huang ◽  
Yanping Li ◽  
Xiaoyi Lv ◽  
Wen Chen ◽  
Shuxian Liu
Keyword(s):  
Feature Selection ◽  
Action Recognition ◽  
Nearest Neighbor ◽  
Health Indicators ◽  
Support Vector ◽  
Normal Walking ◽  
K Nearest Neighbor ◽  
Recognition Algorithms ◽  
Medical Health ◽  
Improved Algorithm

Action recognition algorithms are widely used in the fields of medical health and pedestrian dead reckoning (PDR). The classification and recognition of non-normal walking actions and normal walking actions are very important for improving the accuracy of medical health indicators and PDR steps. Existing motion recognition algorithms focus on the recognition of normal walking actions, and the recognition of non-normal walking actions common to daily life is incomplete or inaccurate, resulting in a low overall recognition accuracy. This paper proposes a microelectromechanical system (MEMS) action recognition method based on Relief-F feature selection and relief-bagging-support vector machine (SVM). Feature selection using the Relief-F algorithm reduces the dimensions by 16 and reduces the optimization time by an average of 9.55 s. Experiments show that the improved algorithm for identifying non-normal walking actions has an accuracy of 96.63%; compared with Decision Tree (DT), it increased by 11.63%; compared with k-nearest neighbor (KNN), it increased by 26.62%; and compared with random forest (RF), it increased by 11.63%. The average Area Under Curve (AUC) of the improved algorithm improved by 0.1143 compared to KNN, by 0.0235 compared to DT, and by 0.04 compared to RF.

Development of a Test Facility and Its Application for Validation and Reliability Testing of Safety-Critical Software

2010 ◽  
Author(s):  
M. Zhao ◽  
N. Tailor
Keyword(s):  
Test Facility ◽  
Actual Process ◽  
Project Schedule ◽  
Human Errors ◽  
Candu Reactors ◽  
Training Tool ◽  
Safety Critical ◽  
Functional Specification ◽  
System A ◽  
Plant Personnel

This paper describes a versatile test facility developed by AECL for validation and reliability (V&R) testing of safety-critical software used in the process trip computers for CANDU reactors. It describes the hardware and software aspects of the test facility. The test hardware consists of a test rig with a test computer used for executing the test software and a process trip computer emulator. The test software is comprised of an operating system, a test interpreter, a test oracle, and a man-machine interface. This paper also discusses the application of the test facility in V&R testing of the process trip computer, how test scripts are prepared and automatically run on the test computer, and how test results are automatically generated by the test computer, thus eliminating potential human errors. The test scripts, which contain specific instructions for testing, are text files written in a special AECL test language. An AECL Test Language Interpreter (ATLIN) program interprets the test scripts and translates structured English statements in the test scripts into test actions. The intuitive nature of the special AECL test language, the version controlled test scripts in text format and automatic test logging feature facilitate the preparation of test cases, which are easy to repeat, review and readily modifiable, and production of consistent results. This paper presents the concept of adding a process trip computer emulator for use in preparation of V&R testing. The process trip computer emulator is designed independently from the actual process trip computer but based on the same functional specification as for the process trip computer. The use of the process trip computer emulator allows the test scripts to be exercised before the actual process trip computers are available for V&R testing, thereby, resulting in a significant improvement to the project schedule. The test facility, with the built-in process trip computer emulator, is also a valuable training tool for the V&R staff and plant personnel.

scholarly journals Marker-Less Motion Capture of Insect Locomotion With Deep Neural Networks Pre-trained on Synthetic Videos

2021 ◽  
Vol 15 ◽  
Author(s):  
Ilja Arent ◽  
Florian P. Schmidt ◽  
Mario Botsch ◽  
Volker Dürr
Keyword(s):  
Motion Capture ◽  
Body Position ◽  
Video Data ◽  
The Other ◽  
Training Data ◽  
Manual Annotation ◽  
Insect Locomotion ◽  
Training Requirements ◽  
Other Hand ◽  
Motion Capture Systems

Motion capture of unrestrained moving animals is a major analytic tool in neuroethology and behavioral physiology. At present, several motion capture methodologies have been developed, all of which have particular limitations regarding experimental application. Whereas marker-based motion capture systems are very robust and easily adjusted to suit different setups, tracked species, or body parts, they cannot be applied in experimental situations where markers obstruct the natural behavior (e.g., when tracking delicate, elastic, and/or sensitive body structures). On the other hand, marker-less motion capture systems typically require setup- and animal-specific adjustments, for example by means of tailored image processing, decision heuristics, and/or machine learning of specific sample data. Among the latter, deep-learning approaches have become very popular because of their applicability to virtually any sample of video data. Nevertheless, concise evaluation of their training requirements has rarely been done, particularly with regard to the transfer of trained networks from one application to another. To address this issue, the present study uses insect locomotion as a showcase example for systematic evaluation of variation and augmentation of the training data. For that, we use artificially generated video sequences with known combinations of observed, real animal postures and randomized body position, orientation, and size. Moreover, we evaluate the generalization ability of networks that have been pre-trained on synthetic videos to video recordings of real walking insects, and estimate the benefit in terms of reduced requirement for manual annotation. We show that tracking performance is affected only little by scaling factors ranging from 0.5 to 1.5. As expected from convolutional networks, the translation of the animal has no effect. On the other hand, we show that sufficient variation of rotation in the training data is essential for performance, and make concise suggestions about how much variation is required. Our results on transfer from synthetic to real videos show that pre-training reduces the amount of necessary manual annotation by about 50%.

scholarly journals Automated measurement of long-term bower behaviors in Lake Malawi cichlids using depth sensing and action recognition

2020 ◽  
Author(s):  
Zachary V Johnson ◽  
Lijiang Long ◽  
Junyu Li ◽  
Manu Tej Sharma Arrojwala ◽  
Vineeth Aljapur ◽  
...  
Keyword(s):  
Decision Making ◽  
Action Recognition ◽  
Lake Malawi ◽  
Simultaneous Recording ◽  
Video Data ◽  
Automated System ◽  
Depth Sensing ◽  
Sand Surface ◽  
In The Wild

ABSTRACTMeasuring naturalistic behaviors in laboratory settings is difficult, and this hinders progress in understanding decision-making in response to ecologically-relevant stimuli. In the wild, many animals manipulate their environment to create architectural constructions, which represent a type of extended phenotype affecting survival and/or reproduction, and these behaviors are excellent models of goal-directed decision-making. Here, we describe an automated system for measuring bower construction in Lake Malawi cichlid fishes, whereby males construct sand structures to attract mates through the accumulated actions of thousands of individual sand manipulation decisions over the course of many days. The system integrates two orthogonal methods, depth sensing and action recognition, to simultaneously measure the developing bower structure and classify the sand manipulation decisions through which it is constructed. We show that action recognition accurately (>85%) classifies ten sand manipulation behaviors across three different species and distinguishes between scooping and spitting events that occur during bower construction versus feeding. Registration of depth and video data streams enables topographical mapping of these behaviors onto a dynamic 3D sand surface. The hardware required for this setup is inexpensive (<$250 per setup), allowing for the simultaneous recording from many independent aquariums. We further show that bower construction behaviors are non-uniform in time, non-uniform in space, and spatially repeatable across trials. We also quantify a unique behavioral phenotype in interspecies hybrids, wherein males sequentially express both phenotypes of behaviorally-divergent parental species. Our work demonstrates that simultaneously tracking both structure and behavior provides an integrated picture of long-term goal-directed decision-making in a naturalistic, dynamic, and social environment.

scholarly journals Action Recognition Using Close-Up of Maximum Activation and ETRI-Activity3D LivingLab Dataset

Sensors ◽  
2021 ◽  
Vol 21 (20) ◽  
pp. 6774
Author(s):  
Doyoung Kim ◽  
Inwoong Lee ◽  
Dohyung Kim ◽  
Sanghoon Lee
Keyword(s):  
Action Recognition ◽  
High Availability ◽  
Video Data ◽  
Experimental Results ◽  
Great Performance ◽  
Action Sequences ◽  
Maximum Activation ◽  
Rich Data ◽  
Video Input ◽  
Dataset Shift

The development of action recognition models has shown great performance on various video datasets. Nevertheless, because there is no rich data on target actions in existing datasets, it is insufficient to perform action recognition applications required by industries. To satisfy this requirement, datasets composed of target actions with high availability have been created, but it is difficult to capture various characteristics in actual environments because video data are generated in a specific environment. In this paper, we introduce a new ETRI-Activity3D-LivingLab dataset, which provides action sequences in actual environments and helps to handle a network generalization issue due to the dataset shift. When the action recognition model is trained on the ETRI-Activity3D and KIST SynADL datasets and evaluated on the ETRI-Activity3D-LivingLab dataset, the performance can be severely degraded because the datasets were captured in different environments domains. To reduce this dataset shift between training and testing datasets, we propose a close-up of maximum activation, which magnifies the most activated part of a video input in detail. In addition, we present various experimental results and analysis that show the dataset shift and demonstrate the effectiveness of the proposed method.

Sign in / Sign up

Export Citation Format

Share Document