Accuracy Of Human Detection Research Articles

Comprehensive multiparametric analysis of human deepfake speech recognition

In this paper, we undertake a novel two-pronged investigation into the human recognition of deepfake speech, addressing critical gaps in existing research. First, we pioneer an evaluation of the impact of prior information on deepfake recognition, setting our work apart by simulating real-world attack scenarios where individuals are not informed in advance of deepfake exposure. This approach simulates the unpredictability of real-world deepfake attacks, providing unprecedented insights into human vulnerability under realistic conditions. Second, we introduce a novel metric to evaluate the quality of deepfake audio. This metric facilitates a deeper exploration into how the quality of deepfake speech influences human detection accuracy. By examining both the effect of prior knowledge about deepfakes and the role of deepfake speech quality, our research reveals the importance of these factors, contributes to understanding human vulnerability to deepfakes, and suggests measures to enhance human detection skills.

Development of a Low-Cost Arduino-Driven Drone for Rapid Search and Rescue in Philippine Terrain

Search and rescue (SAR) operations in the Philippines are often hampered by challenging terrains and the high costs associated with advanced technology. To address these issues, this study developed, tested, and validated a low-cost Arduino-driven drone optimized for SAR missions. The drone integrates an Arduino Mega 2560 microcontroller with an ArduPilot flight controller, GPS module, infrared sensors, ultrasonic sensors, and a thermal camera, alongside communication modules using Bluetooth Low Energy (BLE) and GSM/GPS. The design aimed to create a reliable, efficient tool capable of navigating complex environments, detecting human presence, avoiding obstacles, and maintaining stable communication with ground control. Extensive field testing demonstrated the drone's navigation stability, human detection accuracy of 95%, and obstacle avoidance success rate of 98%. Iterative debugging and optimization further enhanced performance, extending flight time to 30 minutes and improving sensor accuracy and communication reliability. Final validation confirmed the drone's readiness for real-world deployment, proving it to be a cost-effective solution for SAR operations in disasterprone regions. This research highlights the potential of using affordable, readily available components to significantly improve SAR efficiency and effectiveness in the Philippines.

Design and development of artificial intelligence assisted railway gate controlling system using internet of things

A improvised solution for the Indian Railways' human level crossing controls and the attendant reliability issues they cause. In our nation, level crossings have been the scene of several train accidents. So far, there have been zero productive actions. Using an infrared (IR) detector and an audible alert, the proposed system automatically opens and closes level crossing gates based on when trains approach or depart from the crossing. However, it's possible that a car may become stuck between the pedestrian crossing barriers if they're automated. Here, an ultrasonic sensor may identify the obstruction between the crossing gates, and an Internet of Things (IoT) module would inform the train of the situation. When a train approaches or departs from a railroad crossing, infrared sensors detect its presence and alert an Arduino-UNO, which then commands the gates to open or close using a little servomotor stationed nearby. Using the Node MCU and GPS navigational technology, this system also aims to reduce latency in detecting combustion attacks within the compartment. The resulting section of this paper proves the proposed work efficiency by means of graphically illustrates the following metrics such as: human detection accuracy, gate closing efficiency, alerting system efficiency and data storage efficiency.

Comparing job applicant deception in asynchronous vs synchronous video interviews, with and without AI-assisted assessments

PurposeAsynchronous Video Interviews (AVIs) incorporating Artificial Intelligence (AI)-assisted assessment has become popular as a pre-employment screening method. The extent to which applicants engage in deceptive impression management (IM) behaviors during these interviews remains uncertain. Furthermore, the accuracy of human detection in identifying such deceptive IM behaviors is limited. This study seeks to explore differences in deceptive IM behaviors by applicants across video interview modes (AVIs vs Synchronous Video Interviews (SVIs)) and the use of AI-assisted assessment (AI vs non-AI). The study also investigates if video interview modes affect human interviewers' ability to detect deceptive IM behaviors.Design/methodology/approachThe authors conducted a field study with four conditions based on two critical factors: the synchrony of video interviews (AVI vs SVI) and the presence of AI-assisted assessment (AI vs Non-AI): Non-AI-assisted AVIs, AI-assisted AVIs, Non-AI-assisted SVIs and AI-assisted SVIs. The study involved 144 pairs of interviewees and interviewers/assessors. To assess applicants' deceptive IM behaviors, the authors employed a combination of interviewee self-reports and interviewer perceptions.FindingsThe results indicate that AVIs elicited fewer instances of deceptive IM behaviors across all dimensions when compared to SVIs. Furthermore, using AI-assisted assessment in both video interview modes resulted in less extensive image creation than non-AI settings. However, the study revealed that human interviewers had difficulties detecting deceptive IM behaviors regardless of the mode used, except for extensive faking in AVIs.Originality/valueThe study is the first to address the call for research on the impact of video interview modes and AI on interviewee faking and interviewer accuracy. This research enhances the authors’ understanding of the practical implications associated with the use of different video interview modes and AI algorithms in the pre-employment screening process. The study contributes to the existing literature by refining the theoretical model of faking likelihood in employment interviews according to media richness theory and the model of volitional rating behavior based on expectancy theory in the context of AVIs and AI-assisted assessment.

TwSense: Highly Robust Through-the-Wall Human Detection Method Based on COTS Wi-Fi Device

With the popularization of Wi-Fi router devices, the application of device-free sensing has garnered significant attention due to its potential to make our lives more convenient. Wi-Fi signal-based through-the-wall human detection offers practical applications, such as emergency rescue and elderly monitoring. However, the accuracy of through-the-wall human detection is hindered by signal attenuation caused by wall materials and multiple propagation paths of interference. Therefore, through-the-wall human detection presents a substantial challenge. In this paper, we proposed a highly robust through-the-wall human detection method based on a commercial Wi-Fi device (TwSense). To mitigate interference from wall materials and other environmental factors, we employed the robust principal component analysis (OR-PCA) method to extract the target signal of Channel State Information (CSI). Subsequently, we segmented the action-induced Doppler shift feature image using the K-means clustering method. The features of the images were extracted using the Histogram of Oriented Gradients (HOG) algorithm. Finally, these features were fed into an SVM classifier (G-SVM) optimized by a grid search algorithm for action classification and recognition, thereby enhancing human detection accuracy. We evaluated the robustness of the entire system. The experimental results demonstrated that TwSense achieved the highest accuracy of 96%.

Multi-Angle Models and Lightweight Unbiased Decoding-Based Algorithm for Human Pose Estimation

When a top-down method is taken to the task of human pose estimation, the accuracy of joint point localization is often limited by the accuracy of human detection. In addition, conventional algorithms commonly encode the image to generate a heat map before processing, but the systematic error in decoding the heat map back to the original image has an impact on the positioning. Therefore, to address the two problems, we propose an algorithm that uses multiple angle models to generate the human boxes and then performs lightweight decoding to recover the image. The new boxes can better fit humans and the recovery error can be reduced. First, we split the backbone network into three sub-networks, the first sub-network is responsible for generating the original human box, the second sub-network is responsible for generating a coarse pose estimation in the boxes, and the third sub-network is responsible for a high-precision pose estimation. In order to make the human box fit the human body better, with only a small number of interfering pixels inside the box, models of the human boxes with multiple rotation angles are generated. The results from the second sub-network are used to select the best human box. Using this human box as input to the third sub-network can significantly improve the accuracy of the pose estimation. Then to reduce the errors arising from image decoding, we propose a lightweight unbiased decoding strategy that differs from traditional methods by combining multiple possible offsets to select the direction and size of the final offset. On the MPII dataset and the COCO dataset, we compare the proposed algorithm with 11 state-of-the-art algorithms. The experimental results show that the algorithm achieves a large improvement in accuracy for a wide range of image sizes and different metrics.

Domain adaptation assisted automatic real-time human-based video summarization

The video summarization plays a momentous role for real-time surveillance for detection of suspicious human activities in public scenes. However, efficient, and accurate human-centric video summarization in real-time videos is still a challenging issue due to several inconsistencies such as noise, illumination effects, scale, or rotational variance, etc. Though, several prominent research studies are available related to human activities detection but a robust human-centric video summarization approach is a major issue. In this research, we expound an efficient video summary generation model (ReHuSum) that creates a short summary with human. Our approach employs the conception of domain adaptation for efficient summarization via transferring the knowledge of a potent pre-trained benchmark Inception-v3 model. Additionally, a frame differencing redundancy removal technique is adopted that aids in generating a precise and accurate video summary with human. Our model generates efficient short summaries from real-time videos with an overall human detection accuracy of 99.97%. Besides, we present two novel qualitative metrics namely; summarization rate (Summrate) and summarization loss (Summloss) for assessing the efficacy of overall summarization. The approach generates a real-time human-based summary of 46s with a (Summrate) of 98.7% and (Summloss) of 0.013 for an outdoor CCTV with a duration of 25.2 min. Similarly, an indoor CCTV of 40.10 min duration is summarized to 44s at a (Summrate) of 96.5% and (Summloss) of 0.035. The model can be utilized in real-time security applications (i.e., border or home surveillance, critical infrastructure, smart cities, etc.) to generate accurate video summaries with human.

Cross-Perspective Human Behavior Recognition Based on a Joint Sparse Representation and Distributed Adaptation Algorithm Combined with Wireless Optical Transmission

Traditional human behavior recognition needs many training samples. Signal transmission of images and videos via visible light in the body is crucial for detecting specific actions to accelerate behavioral recognition. Joint sparse representation techniques improve identification accuracy by utilizing multi-perspective information, while distributional adaptive techniques enhance robustness by adjusting feature distributions between different perspectives. Combining both techniques enhances recognition accuracy and robustness, enabling efficient behavior recognition in complex environments with multiple perspectives. In this paper, joint sparse representation has been combined with distributed adaptation algorithm to recognize human behavior under the fusion algorithm, and verify the feasibility of the fusion algorithm through experimental analysis. The research objective of this article is to explore the use of the combination of joint sparse representation technology and distributed adaptive technology in the recall and accuracy of human detection, combined with the cross perspective human behavior recognition of wireless optical transmission. The experimental results showed that in the process of human detection, the recall and precision of the fusion algorithm in this paper reached 92% and 90% respectively, which are slightly higher than the comparison algorithm. In the experiment of recognition accuracy of different actions, the recognition accuracy of the fusion algorithm in this paper was also higher than that of the control algorithm. It can be seen that the fusion of joint sparse representation and distributed adaptation algorithms, as well as wireless communication light technology, are of great significance for human behavior recognition.

Implementation of YOLO-v5 for a Real Time Social Distancing Detection

The world is in an uproar with the Covid-19 pandemic, which has had an impact on society. Various efforts have been made by governments around the world to suppress the spread of the Covid-19 virus. One of the health protocols that have been appealed by the government is social distancing or social restrictions, namely limiting interactions between human beings as long as 1-2 meters. But in reality, there are still many people who ignore social distancing policies. The application of a social distancing detection system can be a solution to this problem. This system aims to detect people who are violating health protocols in the form of social distancing and then issue a voice warning to keep their distance from others to avoid the spread of the Covid-19 virus by using the YOLO-v5 method which is the latest version of YOLO (You Only Look Once). . Processing speed of YOLO-v5 has increased drastically with the fastest speed reaching 140 Frames Per Second (FPS) and has a small size, even having a size of 90% compared to the previous version. The accuracy of human detection using YOLO-v5 from this system reaches 83.28% and the accuracy of social distancing detection reaches 90.8%. From the results of the percentage analysis that has been carried out, it can be concluded that the system that has been created can function well for social distancing detection, but it is difficult to detect if humans are too far from the camera.

Development, validation, and relevance of in vivo low-contrast task transfer function to estimate detectability in clinical CT images.

The current state-of-the-art calculation of detectability index (d') is largely phantom-based, with the latest being based on a hybrid phantom noise power spectrum (NPS) combined with patient-specific noise magnitude and high-contrast air-skin interface. The purpose of this study was to develop and assess the use of fully patient-specific measurements of noise and low-contrast resolution, derived entirely from patient images on d'. This study developed a d' calculation that is patient- and task-specific, employing newly developed algorithms for estimating patient-specific NPS and low-contrast task transfer function (TTF). The TTF estimation methodology used a trained regression support vector machine (SVM) to estimate a fitted form of the TTF given a variance-normalized estimate of the NPS (referred to as the TTFNPS ). The regression SVM was trained and tested using five-fold cross-validation on 192 scans (4 dose levels x 6 reconstruction kernels x 4 repeats) of a phantom with low-contrast polyethylene insert and reconstructed with filtered backprojection and iterative reconstructions across 12 clinically relevant kernels (FBP: B20f, B31f, B45f; SAFIRE: I26f, I31f, J45f with strengths: 2, 3, 5). To test the low-contrast TTF estimation method, the estimated TTFNPS measurements were compared to (1) TTF measurements from the air-phantom interface (referred to as the TTFair , representing the most patient-specific clinical alternative) and (2) TTF measurements from the edge of the low-contrast polyethylene insert (referred to as the TTFpoly ), which represented the gold standard of low-contrast TTF measurement. Patient-specific NPS, patient-specific noise magnitude, and patient-specific low-contrast TTF were further combined with a reference task function to calculate a d' (according to a non-prewhitening matched filter model) across 1120 lesions previously evaluated in 2AFC human observer detection of liver lesions. The resulting values were compared to the observer results using a generalized linear mixed-effects statistical model. The correlations between the model and observer results were also compared with previously reported values (using a hybrid method with phantom-derived NPS and TTFair ). The TTFNPS more accurately represented resolution across the considered reconstruction settings, compared with the TTFair . The out-of-fold predictions of the TTFNPS had statistically better root-mean-square error concordance (p<0.05, one-tailed Wilcoxon rank-sum test) to gold standard than the TTFair (the alternative, measured from the air-phantom interface). Detectability indices informed by purely patient-specific NPS and TTF were strongly correlated with 2AFC outcomes (p<0.05). R2 between human detection accuracy and model-predicted detection accuracy were shown to be greater for those measured with patient-specific d' than for the hybrid d' but failed to rise to the level of statistical significance (p≥0.05, bootstrap resampled corrected paired Student's t-test). The results suggest that fully patient-specific characterization of image quality based on in vivo NPS and low-contrast TTF offer advantages over hybrid methods. The results in terms of d' favorably relate to observer detection of liver lesions. The method can potentially be integrated into an automated image quality tracking system to assess image quality across a computed tomography clinical operation without needing phantom scans.

A Video Surveillance System for Unmanned Surveillance of Cantonment Boundary

Since last few years, the Incidents that breach internal security and attack on the security forces are increasing day by day. These are security issues are becoming challenging to handle manually due to economical restrictions. This paper proposes an application for video surveillance to handle and monitor the intrusive incidents. The proposed application includes then human detection in no men’s land around the boundary of the army cantonment. The human detection approach is proposed in this paper is developed with integration of the object detection using background subtraction, feature extraction using CNN and object classification into human and non human using SVM. The proposed approach achieves 95.6% accuracy in human detection. Application proposed in this paper is useful for unmanned surveillance of cantonment boundary.

The left-behind human detection and tracking system based on vision with multi-model fusion and microwave radar inside the bus

Left-behind humans inside the car or bus have caused a lot of accidents, so it is essential to detect the humans in vehicle. Current human detection methods rely on wearable devices, oxygen sensors, and special seat designs in vehicles, but those sensors cannot adapt to ever-changing environments. To solve those problems and especially to improve passengers’ safety on the bus, we propose a method to accomplishing human detection by fusion vision and microwave radar information in various environments in vehicle. For vision information, we use different networks to extract human and human face features, and fusion of the detection results in different models to improve human detection accuracy. The human detection model is MobileNet-V2, and the human face detection model is MTCNN. A new matching schedule and tracking objects management rule based on the Kernelized Correlation Filter tracker are designed to track the human and human face detection boxes. The microwave radar information is used to detect moving objects. Finally, the fusion vision and microwave radar detection results are implemented. Experiments show that our method has improved the human detection accuracy in vehicle, and this method can be used for detection of left-behind children on the school bus.

Audio-Processing-Based Human Detection at Disaster Sites With Unmanned Aerial Vehicle

This paper describes a human search system that uses an unmanned aerial vehicle (UAV). The use of robots to search for people is expected to become an auxiliary tool for saving lives during a disaster. In particular, because UAVs can collect information from the air, there has been much research into human search using UAVs equipped with cameras. However, the disadvantage of cameras is that they struggle to detect people who are hidden in shadows. To solve this problem, we mounted an array microphone on a UAV and to detect the human voice as a means of finding people that cameras cannot. Also a search method is proposed that combines voice and camera human detection to compensate for their respective shortcomings. The rate and accuracy of human detection by the proposed method are assessed experimentally.

HUMAN DETECTION BASED ON A SEQUENCE OF THERMAL IMAGES USING DEEP LEARNING

Abstract. Human detection has been playing an increasingly important role in many fields in recent years. Human detection is a still challenging task because, for the group of people, each individual has his unique appearance and, body shape.Compared with the traditional method, the deep learning neural network has the advantages of shorter computing time, higher accuracy and easier operation. Therefore, deep learning method has been widely used in object detection. The current state of art in human detection is RetinaNet. Among all the deep learning approaches, RetinaNet gives the highest accuracy of human detection (Lin, Goyal, Girshick, He, &amp;amp; Piotr Dollar, 2018).The temporal component of video provides additional and significant clues as compared to the static image. In this paper, the temporal relationship of the images is utilized to improve the accuracy of human detection. Compared to using only an image, the accuracy of human detection is 21.4% higher when a sequence of images is applied.

Edge Computing Framework for Cooperative Video Processing in Multimedia IoT Systems

Multimedia Internet-of-Things (IoT) systems have been widely used in surveillance, automatic behavior analysis and event recognition, which integrate image processing, computer vision, and networking capabilities. In conventional multimedia IoT systems, videos captured by surveillance cameras are required to be delivered to remote IoT servers for video analysis. However, the long-distance transmission of a large volume of video chunks may cause congestions and delays due to limited network bandwidth. Nowadays, mobile devices, e.g., smart phones and tablets, are resource-abundant in computation and communication capabilities. Thus, these devices have the potential to extract features from videos for the remote IoT servers. By sending back only a few video features to the remote servers, the bandwidth starvation of delivering original video chunks can be avoided. In this paper, we propose an edge computing framework to enable cooperative processing on resource-abundant mobile devices for delay-sensitive multimedia IoT tasks. We identify that the key challenges in the proposed edge computing framework are to optimally form mobile devices into video processing groups and to dispatch video chunks to proper video processing groups. Based on the derived optimal matching theorem, we put forward a cooperative video processing scheme formed by two efficient algorithms to tackle above challenges, which achieves suboptimal performance on the human detection accuracy. The proposed scheme has been evaluated under diverse parameter settings. Extensive simulation confirms the superiority of the proposed scheme over other two baseline schemes.

CSI‐based low‐duty‐cycle wireless multimedia sensor network for security monitoring

Wireless multimedia sensor networks (WMSNs) have been proposed for use in many challenging applications, such as military surveillance, wildlife tracking, and security monitoring. Reducing energy and storage space consumption are challenging in WMSNs. To address this issue, the authors propose a novel low-duty-cycle WMSN (LWMSN) based on Wi-Fi. This special WMSN is especially suitable for security monitoring, which involves long invalid times in which humans do not appear in the target area. The LWMSN can be intelligently switched on and off by human detection based on channel state information provided by Wi-Fi products. The authors evaluated their design with a simulation and an experiment. Results show that the accuracy of human detection can be >97% on average. In addition, energy and storage space consumption are reduced by 53 and 40%, respectively.

Estimating detectability index in vivo: development and validation of an automated methodology

This study's purpose was to develop and validate a method to estimate patient-specific detectability indices directly from patients' CT images (i.e., in vivo). The method extracts noise power spectrum (NPS) and modulation transfer function (MTF) resolution properties from each patient's CT series based on previously validated techniques. These are combined with a reference task function (10-mm disk lesion with [Formula: see text] HU contrast) to estimate detectability indices for a nonprewhitening matched filter observer model. This method was applied to CT data from a previous study in which diagnostic performance of 16 readers was measured for the task of detecting subtle, hypoattenuating liver lesions ([Formula: see text]), using a two-alternative-forced-choice (2AFC) method, over six dose levels and two reconstruction algorithms. In vivo detectability indices were estimated and compared to the human readers' binary 2AFC outcomes using a generalized linear mixed-effects statistical model. The results of this modeling showed that the in vivo detectability indices were strongly related to 2AFC outcomes ([Formula: see text]). Linear comparison between human-detection accuracy and model-predicted detection accuracy (for like conditions) resulted in Pearson and Spearman correlation coefficients exceeding 0.84. These results suggest the potential utility of using in vivo estimates of a detectability index for an automated image quality tracking system that could be implemented clinically.

Human Detection System by Fusing Depth Map-Based Method and Convolutional Neural Network-Based Method

In this paper, the depth images and the colour images provided by Kinect sensors are used to enhance the accuracy of human detection. The depth-based human detection method is fast but less accurate. On the other hand, the faster region convolutional neural network-based human detection method is accurate but requires a rather complex hardware configuration. To simultaneously leverage the advantages and relieve the drawbacks of each method, one master and one client system is proposed. The final goal is to make a novel Robot Operation System (ROS)-based Perception Sensor Network (PSN) system, which is more accurate and ready for the real time application. The experimental results demonstrate the outperforming of the proposed method compared with other conventional methods in the challenging scenarios.

Indoor Human Detection Based on Thermal Array Sensor Data and Adaptive Background Estimation

Low Resolution Thermal Array Sensors are widely used in several applications in indoor environments. In particular, one of these cheap, small and unobtrusive sensors provides a low-resolution thermal image of the environment and, unlike cameras; it is capable to detect human heat emission even in dark rooms. The obtained thermal data can be used to monitor older seniors while they are performing daily activities at home, to detect critical situations such as falls. Most of the studies in activity recognition using Thermal Array Sensors require human detection techniques to recognize humans passing in the sensor field of view. This paper aims to improve the accuracy of the algorithms used so far by considering the temperature environment variation. This method leverages an adaptive background estimation and a noise removal technique based on Kalman Filter. In order to properly validate the system, a novel installation of a single sensor has been implemented in a smart environment: the obtained results show an improvement in human detection accuracy with respect to the state of the art, especially in case of disturbed environments.

Correlation between human detection accuracy and observer model-based image quality metrics in computed tomography.

The purpose of this study was to compare computed tomography (CT) low-contrast detectability from human readers with observer model-based surrogates of image quality. A phantom with a range of low-contrast signals (five contrasts, three sizes) was imaged on a state-of-the-art CT scanner (Siemens' force). Images were reconstructed using filtered back projection and advanced modeled iterative reconstruction and were assessed by 11 readers using a two alternative forced choice method. Concurrently, contrast-to-noise ratio (CNR), area-weighted CNR (CNRA), and observer model-based metrics were estimated, including nonprewhitening (NPW) matched filter, NPW with eye filter (NPWE), NPW with internal noise, NPW with an eye filter and internal noise (NPWEi), channelized Hotelling observer (CHO), and CHO with internal noise (CHOi). The correlation coefficients (Pearson and Spearman), linear discriminator error, [Formula: see text], and magnitude of confidence intervals, [Formula: see text], were used to determine correlation, proper characterization of the reconstruction algorithms, and model precision, respectively. Pearson (Spearman) correlation was 0.36 (0.33), 0.83 (0.84), 0.84 (0.86), 0.86 (0.88), 0.86 (0.91), 0.88 (0.90), 0.85 (0.89), and 0.87 (0.84), [Formula: see text] was 0.25, 0.15, 0.2, 0.25, 0.3, 0.25, 0.4, and 0.45, and [Formula: see text] was [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text] for CNR, CNRA, NPW, NPWE, NPWi, NPWEi, CHO, and CHOi, respectively.