Search

Featured Application: This work presents the use of device-free wireless sensing for crowd size estimation on subway platforms. Dense urban environments pose significant challenges when it comes to detecting and measuring crowd size due to their nature of being free-flow environments containing many dynamic factors. In this paper, we use a wireless sensor network (WSN) to perform device-free crowd size estimation in a subway station. Our sensing solution uses the change in attenuation of the communication links between sensor nodes to estimate the number of people standing on the platform. In order to achieve this, we use the same attenuation information coming from the WSN to detect the presence of a rail vehicle in the station and compensate for the channel fading caused by the introduced rail vehicle. We make use of two separately trained regression models depending on the presence or absence of a rail vehicle to estimate the people count. The detection of rail vehicles occurred with a near-perfect accuracy. When evaluating the resulting estimation model on our test set, we achieved a mean average error of 3.567 people, which is a significant improvement over 6.192 people when using a single regression model. This demonstrates that device-free sensing technologies can be successfully implemented in dynamic environments by implementing detection techniques and using different regression models depending on the environment's state. [ABSTRACT FROM AUTHOR]

The early detection of the acute deterioration of escalating illness severity is crucial for effective patient management and can significantly impact patient outcomes. Ambient sensing technology, such as computer vision, may provide real-time information that could impact early recognition and response. This study aimed to develop a computer vision model to quantify the number and type (clinician vs. visitor) of people in an intensive care unit (ICU) room, study the trajectory of their movement, and preliminarily explore its relationship with delirium as a marker of illness severity. To quantify the number of people present, we implemented a counting-by-detection supervised strategy using images from ICU rooms. This was accomplished through developing three methods: single-frame, multi-frame, and tracking-to-count. We then explored how the type of person and distribution in the room corresponded to the presence of delirium. Our designed pipeline was tested with a different set of detection models. We report model performance statistics and preliminary insights into the relationship between the number and type of persons in the ICU room and delirium. We evaluated our method and compared it with other approaches, including density estimation, counting by detection, regression methods, and their adaptability to ICU environments. [ABSTRACT FROM AUTHOR]

Indoor occupancy detection (IOD) via Wi-Fi sensing capitalizes on the varying patterns in CSI (Channel State Information) to estimate the number of people in a given area. However, the precision of such systems heavily depends on the quality of the CSI data, which can be degraded by noise and environmental factors. To address this issue, In this paper, we present a CSI preprocessing method to improve the accuracy of IOD systems using Wi-Fi sensing. Unlike existing preprocessing methods that use computationally complex signal processing or statistical techniques, we expand the dimension of CSI amplitude data into a three-channel vector through nonlinear transformation to amplify subtle differences between CSI data belonging to a different number of people. By drawing clearer boundaries between CSI data distributions belonging to a different number of people in a monitored area, our method improves the people-counting accuracy of a Wi-Fi sensing system. To ensure temporal consistency and improve data quality, we discretize the CSI measurements based on their transmission periods and aggregate consecutive measurements over a given time interval. These samples are then fed into a Convolutional Neural Network (CNN) specifically trained for the IOD task. Experimental results in diverse real-world scenarios verify that compared to the traditional methods, the enhanced feature representation capability of our approach leads to more accurate and robust sensing outcomes even in the most resource-constrained environment, where a commercial off-the-shelf CSI capture machine with only one antenna is used when a Wi-Fi sender with one transmit antenna sends packets periodically to the channel with the smallest Wi-Fi channel bandwidth. [ABSTRACT FROM AUTHOR]

Occupancy control is a fundamental aspect of managing spaces and services effectively. It aims to ensure safety, compliance with regulations, emergency preparedness, and overall satisfaction for individuals and businesses. To align with the described need, this paper presents a computer vision based system for automatic people counting in gates. The system is divided in five stages: video capture, motion analysis, human detection, human tracking and people counting. A camera captures the top-view image of the gate and analyze the change or movement in the objects in scene. When motion is detected, the frame is sent to the object detector, which is a convolutional neural network. Then, a tracking algorithm analyzes the movement patterns of people. According to the route, it is determined whether the person arrives or leaves and the count is updated. Two test scenarios are analyzed: the entry of a public bus and a building gate. The people detection module is tested, showing a mean average precision of 95.2%. Also, the counting is tested showing an average precision of 96.8%, a recall of 92% and an F1-Score of 94.3%. Finally, the system performance is evaluated, showing an average processing time of 34.2 ms. [ABSTRACT FROM AUTHOR]

Traditional crowd counting (optical flow or feature matching) techniques have been upgraded to deep learning (DL) models due to their lack of automatic feature extraction and low-precision outcomes. Most of these models were tested on surveillance scene crowd datasets captured by stationary shooting equipment. It is very challenging to perform people counting from the videos shot with a head-mounted moving camera; this is mainly due to mixing the temporal information of the moving crowd with the induced camera motion. This study proposed a transfer learning-based PeopleNet model to tackle this significant problem. For this, we have made some significant changes to the standard VGG16 model, by disabling top convolutional blocks and replacing its standard fully connected layers with some new fully connected and dense layers. The strong transfer learning capability of the VGG16 network yields in-depth insights of the PeopleNet into the good quality of density maps resulting in highly accurate crowd estimation. The performance of the proposed model has been tested over a self-generated image database prepared from moving camera video clips, as there is no public and benchmark dataset for this work. The proposed framework has given promising results on various crowd categories such as dense, sparse, average, etc. To ensure versatility, we have done self and cross-evaluation on various crowd counting models and datasets, which proves the importance of the PeopleNet model in adverse defense of society. [ABSTRACT FROM AUTHOR]

Occupancy control is a fundamental aspect of managing spaces and services effectively. It aims to ensure safety, compliance with regulations, emergency preparedness, and overall satisfaction for individuals and businesses. To align with the described need, this paper presents a computer vision based system for automatic people counting in gates. The system is divided in five stages: video capture, motion analysis, human detection, human tracking and people counting. An RGB camera captures the top-view image of the gate and analyze the change or movement in the objects in scene. When motion is detected, the frame is sent to the object detector, which is a convolutional neural network. Then, a tracking algorithm analyzes the movement patterns of people. According to the route, it is determined whether the person arrives or leaves and the count is updated. Two test scenarios are analyzed: the entry of a public bus and a building gate. The people detection module is tested, showing a mAP of 95.2% and a mean IoU (50%) of 55.9%. Also, the counting is tested showing an average precision of 96.8%, a recall of 92% and an F1-Score of 94.3%. Finally, the system performance is evaluated, showing an average processing time of 34.2 ms and 29.2 FPS.

Monitoring the number of people in various spaces of a building is important for optimizing space usage, assisting with public safety, and saving energy. Diverse approaches have been developed for different end goals, from ID card readers for space management, to surveillance cameras for security, to CO2 sensing for HVAC control. In the last few years, fisheye cameras mounted overhead have become the sensing modality of choice because they offer large-area coverage and significantly-reduced occlusions but research efforts are still nascent. In this paper, we provide an overview of recent research efforts in this area and propose one new direction. First, we identify benefits and challenges related to inference from top-view fisheye images, and summarize key public datasets. Then, we review efforts in algorithm development for detecting people from a single fisheye frame and from a group of sequential frames. Finally, we focus on counting people indoors. While this is straightforward for a single camera, when multiple cameras are used to monitor a space, person re-identification is needed to avoid overcounting. We describe a framework for people counting using two cameras and demonstrate its effectiveness in a large classroom for location-based person re-identification. To support people counting in even larger spaces, we propose two new person re-identification algorithms using N > 2 overhead fisheye cameras. We provide ample experimental results throughout the paper.

Buildings are rapidly becoming more digitized, largely due to developments in the internet of things (IoT). This provides both opportunities and challenges. One of the central challenges in the process of digitizing buildings is the ability to monitor these buildings' status effectively. This monitoring is essential for services that rely on information about the presence and activities of individuals within different areas of these buildings. Occupancy information (including people counting, occupancy detection, location tracking, and activity detection) plays a vital role in the management of smart buildings. In this article, we primarily focus on the use of passive infrared (PIR) sensors for gathering occupancy information. PIR sensors are among the most widely used sensors for this purpose due to their consideration of privacy concerns, cost-effectiveness, and low processing complexity compared to other sensors. Despite numerous literature reviews in the field of occupancy information, there is currently no literature review dedicated to occupancy information derived specifically from PIR sensors. Therefore, this review analyzes articles that specifically explore the application of PIR sensors for obtaining occupancy information. It provides a comprehensive literature review of PIR sensor technology from 2015 to 2023, focusing on applications in people counting, activity detection, and localization (tracking and location). It consolidates findings from articles that have explored and enhanced the capabilities of PIR sensors in these interconnected domains. This review thoroughly examines the application of various techniques, machine learning algorithms, and configurations for PIR sensors in indoor building environments, emphasizing not only the data processing aspects but also their advantages, limitations, and efficacy in producing accurate occupancy information. These developments are crucial for improving building management systems in terms of energy efficiency, security, and user comfort, among other operational aspects. The article seeks to offer a thorough analysis of the present state and potential future advancements of PIR sensor technology in efficiently monitoring and understanding occupancy information by classifying and analyzing improvements in these domains. [ABSTRACT FROM AUTHOR]

Since the emergence of big data, the popularity of deep learning models has increased and they are being implemented in a wide range of applications, including people detection and counting in congested environments. Detecting and counting people for human behavior analysis in retail stores is a challenging research problem due to the congested and crowded environment. This paper proposes a deep learning approach for detecting and counting people in the presence of occlusions and illuminance variation in a crowded retail environment, utilizing deep CNNs (DCNNs) for semantic segmentation of top-view depth visual data. Semantic segmentation has been implemented using (DCNNs) in recent years since it is a powerful approach. The objective of this paper is to design a novel architecture that consists of an encoder–decoder architecture. We were motivated to use transfer learning to solve the problem of insufficient training data. We used ResNet50 for the encoder, and we built the decoder part as a novel contribution. Our model was trained and evaluated on the TVHeads dataset and the people counting dataset (PCDS) that are available for research purposes. It consists of depth data of people captured from a top-view RGB-D sensor. The segmentation results indicate high accuracy and demonstrate that the proposed model is robust and accurate. [ABSTRACT FROM AUTHOR]

Dense urban environments pose significant challenges when it comes to detecting and measuring crowd size due to their nature of being free-flow environments containing many dynamic factors. In this paper, we use a wireless sensor network (WSN) to perform device-free crowd size estimation in a subway station. Our sensing solution uses the change in attenuation of the communication links between sensor nodes to estimate the number of people standing on the platform. In order to achieve this, we use the same attenuation information coming from the WSN to detect the presence of a rail vehicle in the station and compensate for the channel fading caused by the introduced rail vehicle. We make use of two separately trained regression models depending on the presence or absence of a rail vehicle to estimate the people count. The detection of rail vehicles occurred with a near-perfect accuracy. When evaluating the resulting estimation model on our test set, we achieved a mean average error of 3.567 people, which is a significant improvement over 6.192 people when using a single regression model. This demonstrates that device-free sensing technologies can be successfully implemented in dynamic environments by implementing detection techniques and using different regression models depending on the environment’s state.

The early detection of the acute deterioration of escalating illness severity is crucial for effective patient management and can significantly impact patient outcomes. Ambient sensing technology, such as computer vision, may provide real-time information that could impact early recognition and response. This study aimed to develop a computer vision model to quantify the number and type (clinician vs. visitor) of people in an intensive care unit (ICU) room, study the trajectory of their movement, and preliminarily explore its relationship with delirium as a marker of illness severity. To quantify the number of people present, we implemented a counting-by-detection supervised strategy using images from ICU rooms. This was accomplished through developing three methods: single-frame, multi-frame, and tracking-to-count. We then explored how the type of person and distribution in the room corresponded to the presence of delirium. Our designed pipeline was tested with a different set of detection models. We report model performance statistics and preliminary insights into the relationship between the number and type of persons in the ICU room and delirium. We evaluated our method and compared it with other approaches, including density estimation, counting by detection, regression methods, and their adaptability to ICU environments.

Indoor occupancy detection (IOD) via Wi-Fi sensing capitalizes on the varying patterns in CSI (Channel State Information) to estimate the number of people in a given area. However, the precision of such systems heavily depends on the quality of the CSI data, which can be degraded by noise and environmental factors. To address this issue, In this paper, we present a CSI preprocessing method to improve the accuracy of IOD systems using Wi-Fi sensing. Unlike existing preprocessing methods that use computationally complex signal processing or statistical techniques, we expand the dimension of CSI amplitude data into a three-channel vector through nonlinear transformation to amplify subtle differences between CSI data belonging to a different number of people. By drawing clearer boundaries between CSI data distributions belonging to a different number of people in a monitored area, our method improves the people-counting accuracy of a Wi-Fi sensing system. To ensure temporal consistency and improve data quality, we discretize the CSI measurements based on their transmission periods and aggregate consecutive measurements over a given time interval. These samples are then fed into a Convolutional Neural Network (CNN) specifically trained for the IOD task. Experimental results in diverse real-world scenarios verify that compared to the traditional methods, the enhanced feature representation capability of our approach leads to more accurate and robust sensing outcomes even in the most resource-constrained environment, where a commercial off-the-shelf CSI capture machine with only one antenna is used when a Wi-Fi sender with one transmit antenna sends packets periodically to the channel with the smallest Wi-Fi channel bandwidth.

In many industrial or healthcare contexts, keeping track of the number of people is essential. Radar systems, with their low overall cost and power consumption, enable privacy-friendly monitoring in many use cases. Yet, radar data are hard to interpret and incompatible with most computer vision strategies. Many current deep learning-based systems achieve high monitoring performance but are strongly context-dependent. In this work, we show how context generalization approaches can let the monitoring system fit unseen radar scenarios without adaptation steps. We collect data via a 60 GHz frequency-modulated continuous wave in three office rooms with up to three people and preprocess them in the frequency domain. Then, using meta learning, specifically the Weighting-Injection Net, we generate relationship scores between the few training datasets and query data. We further present an optimization-based approach coupled with weighting networks that can increase the training stability when only very few training examples are available. Finally, we use pool-based sampling active learning to fine-tune the model in new scenarios, labeling only the most uncertain data. Without adaptation needs, we achieve over 80% and 70% accuracy by testing the meta learning algorithms in new radar positions and a new office, respectively. [ABSTRACT FROM AUTHOR]

Due to the outstanding penetrating detection performance of low-frequency electromagnetic waves, through-wall radar (TWR) has gained widespread applications in various fields, including public safety, counterterrorism operations, and disaster rescue. TWR is required to accomplish various tasks, such as people detection, people counting, and positioning in practical applications. However, most current research primarily focuses on one or two tasks. In this paper, we propose a multitask network that can simultaneously realize people counting, action recognition, and localization. We take the range–time–Doppler (RTD) spectra obtained from one-dimensional (1D) radar signals as datasets and convert the information related to the number, motion, and location of people into confidence matrices as labels. The convolutional layers and novel attention modules automatically extract deep features from the data and output the number, motion category, and localization results of people. We define the total loss function as the sum of individual task loss functions. Through the loss function, we transform the positioning problem into a multilabel classification problem, where a certain position in the distance confidence matrix represents a certain label. On the test set consisting of 10,032 samples from through-wall scenarios with a 24 cm thick brick wall, the accuracy of people counting can reach 96.94%, and the accuracy of motion recognition is 96.03%, with an average distance error of 0.12 m. [ABSTRACT FROM AUTHOR]

As the Artificial Intelligence of Things (AIOT) and ubiquitous sensing technologies have been leaping forward, numerous scholars have placed a greater focus on the use of Impulse Radio Ultra-Wide Band (IR-UWB) radar signals for Region of Interest (ROI) population estimation. To address the problem concerning the fact that existing algorithms or models cannot accurately detect the number of people counted in ROI from low signal-to-noise ratio (SNR) received signals, an effective 1DCNN-LSTM model was proposed in this study to accurately detect the number of targets even in low-SNR environments with considerable people. First, human-induced excess kurtosis was detected by setting a threshold using the optimized CLEAN algorithm. Next, the preprocessed IR-UWB radar signal pulses were bundled into frames, and the resulting peaks were grouped to develop feature vectors. Subsequently, the sample set was trained based on the 1DCNN-LSTM algorithm neural network structure. In this study, the IR-UWB radar signal data were acquired from different real environments with different numbers of subjects (0–10). As indicated by the experimental results, the average accuracy of the proposed 1DCNN-LSTM model for the recognition of people counting reached 86.66% at ROI. In general, a high-accuracy, low-complexity, and high-robustness solution in IR-UWB radar people counting was presented in this study. [ABSTRACT FROM AUTHOR]

The use of higher frequency bands compared to other wireless communication protocols enhances the capability of accurately determining locations from ultra-wideband (UWB) signals. It can also be used to estimate the number of people in a room based on the waveform of the channel impulse response (CIR) from UWB transceivers. In this paper, we apply deep neural networks to UWB CIR signals for the purpose of estimating the number of people in a room. We especially focus on empirically investigating the various network architectures for classification from single UWB CIR data, as well as from various ensemble configurations. We present our processes for acquiring and preprocessing CIR data, our designs of the different network architectures and ensembles that were applied, and the comparative experimental evaluations. We demonstrate that deep neural networks can accurately classify the number of people within a Line of Sight (LoS), thereby achieving an 99% performance and efficiency with respect to both memory size and FLOPs (Floating Point Operations Per Second). [ABSTRACT FROM AUTHOR]

When a safety accident occurs in a coal mine, it is necessary to clearly understand the personnel situation in each area and reasonably arrange the rescue plan. In this paper, YOLOv5 is used as the target detector, combined with the improved DeepSORT tracking algorithm to track the mine personnel, and the personnel count in each roadway area of the mine is realized. Firstly, the lightweight full-scale feature learning Re-ID feature extraction model OSNet is used to optimize DeepSORT and replace the original CNN feature extraction module. Then, the strategy of training the detector and OSNet feature extraction model separately is adopted to achieve a stable tracking effect in the complex environment of the mine. On this basis, the ROI area and baseline are set in the video screen to judge the situation of people entering and leaving, so as to realize the counting function. In order to effectively train and evaluate the performance of the model, 10 000 pictures of different areas of various roadways under coal mines were collected for training and testing. The MOTA of the improved model was 66.7%, better than that of the former 63.4%. The improved speed is 28.1 FPS, which is better than the 25.3 FPS before the improvement. The experimental results show that the improved model can effectively count mine personnel and can be used in the actual production environment.

Person re-identification (PRID) has been thoroughly researched in typical surveillance scenarios where various scenes are monitored by side-mounted, rectilinear-lens cameras. To date, few methods have been proposed for fisheye cameras mounted overhead and their performance is lacking. In order to close this performance gap, we propose a multi-feature framework for fisheye PRID where we combine deep-learning, color-based and location-based features by means of novel feature fusion. We evaluate the performance of our framework for various feature combinations on FRIDA, a public fisheye PRID dataset. The results demonstrate that our multi-feature approach outperforms recent appearance-based deep-learning methods by almost 18% points and location-based methods by almost 3% points in matching accuracy. We also demonstrate the potential application of the proposed PRID framework to people counting in large, crowded indoor spaces.

The emergence of automation in the current economic trend promotes the usage of computer vision systems in various applications. Counting people in a specified area or on the street can bring many benefits in terms of security and marketing. The people counting system is one of the applications that utilize the computer vision system to count people with higher reliability and accuracy. Thus, this project is to develop an offline automated people counting system based on captured video file input using MATLAB software and a notification system to update and send notifications about the number of occupants in a target area using ThingSpeak. For project development, simulation and development of coding for object detection that involves deep learning approach, object tracking and counting, and development of notification system have been done. Three videos were taken to be used for three trials to evaluate the functionality and performance of the developed system. Based on the results and analysis, the system can perform people detection, people tracking and people counting on the recorded input videos with high accuracy of 94.45%, visualize the data on the ThingSpeak platform and send notifications through Twitter. [ABSTRACT FROM AUTHOR]

Background: Intelligent conference rooms are crucial to 21st-century enterprises for events. Safety, resource optimization, and event management depend on accurate counting in such contexts. Manual headcounts are effective yet inefficient and error-prone, particularly for big crowds, requiring automatic people counters. Objective: This article introduces and validates a data-driven algorithm to count and track people in an intelligent conference hall. The concept uses IoT infrastructure, low-resolution cameras, and powerful image-processing algorithms to improve security, resource usage, and real-time management choices. Methods: The message-oriented IoT algorithm incorporates motion detection, background subtraction, people counting, and tracking modules. Blob analysis, edge detection, and low-maintenance, low-resolution cameras are used to capture real-world data. Based on real-time data, a decision-making module controls the conference hall's atmosphere. Results: With a 96.5% accuracy rate and 95% confidence interval in real-time individual counts, the algorithm operates with exceptional dependability. Using real-world data and experimental findings, the algorithm has been extensively tested and shown to work in diverse head counting situations. Conclusion: Intelligent conference hall management using the suggested algorithm might revolutionize venue management. The algorithm's accurate, real-time headcounts improve security, resource utilization, and management decisions, making it a promising candidate for intelligent conference hall management and optimization for diverse events and gatherings.

As seen in the COVID-19 pandemic, one of the most important measures is physical distance in viruses transmitted from person to person. According to the World Health Organization (WHO), it is mandatory to have a limited number of people in indoor spaces. Depending on the size of the indoors, the number of persons that can fit in that area varies. Then, the size of the indoor area should be measured and the maximum number of people should be calculated accordingly. Computers can be used to ensure the correct application of the capacity rule in indoors monitored by cameras. In this study, a method is proposed to measure the size of a prespecified region in the video and count the people there in real time. According to this method: (1) predetermining the borders of a region on the video, (2) identification and counting of people in this specified region, (3) it is aimed to estimate the size of the specified area and to find the maximum number of people it can take. For this purpose, the You Only Look Once (YOLO) object detection model was used. In addition, Microsoft COCO dataset pre-trained weights were used to identify and label persons. YOLO models were tested separately in the proposed method and their performances were analyzed. Mean average precision (mAP), frame per second (fps), and accuracy rate metrics were found for the detection of persons in the specified region. While the YOLO v3 model achieved the highest value in accuracy rate and mAP (both 0.50 and 0.75) metrics, the YOLO v5s model achieved the highest fps rate among non-Tiny models. [ABSTRACT FROM AUTHOR]

In recent years, multiple IoT solutions have used computational intelligence technologies to identify people and count them. WIFI Channel State Information (CSI) has recently been applied to counting people with multiple benefits, such as being cost-effective, easily accessible, free of privacy concerns, etc. However, most current CSI-based work is limited to human location-fixed environments since human location-random environments are more complicated. Aiming to fix the problem of counting people in human location-random environments, we propose a solution using deep learning CM-NET, an end-to-end cross-modal learning network. Since it is difficult to count people with CSI straightforwardly, CM-NET approaches this problem using deep learning, utilizing a multi-layer transformer model to automatically extract the correlations between channels and the number of people. Owing to the complexity of human location-random environments, the transformer model cannot extract characteristics describing the number of people. To enhance the feature learning capability of the transformer model, CM-NET takes the feature knowledge learned by the image-based people counting model to supervise the learning process. In particular, CM-NET works with CSI alone during the testing phase without any image information, and ultimately achieves sound results with an average accuracy of 86%. Meanwhile, the superiority of CM-NET has been verified by comparison with the latest available related methods. [ABSTRACT FROM AUTHOR]

HRP UWB technology has recently attracted attention owing to its advantages of enabling precise distance measurements as well as communication. In addition to communication and distance measurement, passive radar applications that can recognize the surrounding situation are also being considered. Among these radar applications, the ability to detect the number of people indoors is of increasing interest when social distancing is important. In this study, we implemented an HRP UWB passive radar application system that detects the number of people in an indoor environment using the change in channel characteristics between HRP UWB communication devices. Using the implemented system, changes in the channel impulse response (CIR) characteristics of channels 5 and 9 were observed, and the number of people was identified using the singular vector decomposition (SVD) algorithm. From actual measurements, it was possible to confirm that the number of people within five people in an indoor environment could be accurately identified using a pair of HRP UWB wireless communication devices in an indoor environment. [ABSTRACT FROM AUTHOR]

In recent years, the performance of people-counting models has been dramatically increased that they can be implemented in practical cases. However, the current models can only count all of the people captured in the inputted closed circuit television (CCTV) footage. Oftentimes, we only want to count people in a specific Region-of-Interest (RoI) in the footage. Unfortunately, simple approaches such as covering the area outside of the RoI are not applicable without degrading the performance of the models. Therefore, we developed a novel learning strategy that enables a deep-learning-based people counting model to count people only in a certain RoI. In the proposed method, the people counting model has two heads that are attached on top of a crowd counting backbone network. These two heads respectively learn to count people inside the RoI and negate the people count outside the RoI. We named this proposed method Gap Regularizer and tested it on ResNet-50, ResNet-101, CSRNet, and SFCN. The experiment results showed that Gap Regularizer can reduce the mean absolute error (MAE), root mean square error (RMSE), and grid average mean error (GAME) of ResNet-50, which is the smallest CNN model, with the highest reduction of 45.2%, 41.25%, and 46.43%, respectively. On shallow models such as the CSRNet, the regularizer can also drastically increase the SSIM by up to 248.65% in addition to reducing the MAE, RMSE, and GAME. The Gap Regularizer can also improve the performance of SFCN which is a deep CNN model with back-end features by up to 17.22% and 10.54% compared to its standard version. Moreover, the impacts of the Gap Regularizer on these two models are also generally statistically significant (P-value < 0.05) on the MOT17-09, MOT20-02, and RHC datasets. However, it has a limitation in which it is unable to make significant impacts on deep models without back-end features such as the ResNet-101.

Pandemi Covid-19 saat ini cukup memberikan dampak pada masyarakat. Skala penyebaran dari Covid-19 sangatlah cepat, sehingga membutuhkan penanganan yang benar. Salah satu cara untuk mengurangi penyebaran Covid-19 adalah dengan melakukan Social Distancing. Namun masyarakat cenderung lalai dalam melaksanakan protokol kesehatan tersebut. Salah satu cara untuk mengatasi masalah ini adalah dengan aplikasi social distancing detector yaitu aplikasi yang digunakan untuk mendeteksi jumlah dan jarak dari objek manusia yang ada pada satu area. Penelitian ini bertujuan untuk mengembangkan aplikasi social distancing detector menggunakan bahasa pemrograman Python dengan library YOLOv3. YOLOv3 memiliki kelebihan dalam object detection dengan akurasi yang tinggi yaitu diatas 90%. Pengujian metode pada penelitian ini menggunakan lima dataset pejalan kaki dari kamera pengawas jalan yang didapatkan dari dataset uji coba beberapa peneliti melalui Github yang memiliki resolusi yang baik dan memiliki objek manusia yang heterogen. Hasil akurasi dari deteksi citra pertama adalah 83,32%. Hasil akurasi dari deteksi citra kedua adalah 76,92%. Hasil akurasi dari deteksi citra ketiga adalah 89,99%. Hasil akurasi dari deteksi citra keempat dan kelima adalah 100%. Hasil Rata-rata tingkat keberhasilan dari semua hasil analisa adalah 90,04% yang diukur dari rata-rata perbandingan jumlah data percobaan berhasil dan jumlah data pengamatan untuk tiap-tiap citra.

A number of occupancy-sensing methods have been proposed to date with the goal of saving HVAC energy in commercial spaces by means of demand-controlled ventilation. Currently, CO 2 measurement is the most common approach deployed in practice. However, recently a number of other approaches have been proposed, such as using surveillance cameras, depth sensors, thermal imagers, etc. In this paper, we compare the occupancy-sensing performance of high-accuracy commercial CO 2 sensors and overhead fisheye cameras supported by deep-learning algorithms. Our experiments are conducted over 3 days in a large university classroom with highly-dynamic occupancy. First, we assess the impact of parameter selection and sensor placement on CO 2 -to-occupancy conversion accuracy. Then, for the best-performing setup, we compare this accuracy against that of fisheye cameras. Subsequently, we estimate the potential energy savings offered by both systems. Our results show that overhead fisheye cameras produce on average a 20% lower occupancy-estimation error than CO 2 sensors in steady-state periods, and almost 70% lower error in transient periods across 4 performance metrics used in this work. In terms of potential energy savings, fisheye cameras offer on average 6 percentage points of additional savings compared to CO 2 sensors. Given their relatively low cost and implementation simplicity, CO 2 sensors are an attractive approach to saving HVAC energy, however their occupancy-estimation accuracy is highly sensitive to parameter selection that is challenging in practice. Occupancy sensing based on fisheye cameras is not subject to similar parameterization sensitivity and in addition to outperforming CO 2 sensors allows precise localization of occupants in a space thus potentially enabling additional space-management and safety/security services, that CO 2 sensors cannot offer. [ABSTRACT FROM AUTHOR]

Handling and interpreting sparse 3D point clouds, especially from mmWave radar, presents unique challenges due to the inherent data sparsity and the vast domain difference compared to denser point clouds like those from LiDAR. In this paper, we introduce a novel cascaded generative adversarial network (GAN) approach to bridge this domain gap. The core principle is to progressively refine the radar-based point cloud through a series of GANs, each targeting a higher resolution. By leveraging multi-level features and a hybrid loss function that combines adversarial, geometric, and consistency components, our method ensures a smooth transition from the sparse radar representation to a high-resolution LiDAR-like point cloud. Our cascaded approach operates at a patch level, and the integrated loss function ensures that the generated points not only resemble the target domain but also maintain geometric and structural fidelity. Real-life dataset consisting mostly of moving pedestrians were collected using a system made of Radar, LiDAR, and RGB Camera. Through an extensive experiment on the collected real-world pedestrian dataset, we validate the efficacy of our approach. Inference from the network indicates that our method can upsample mmWave radar point clouds with enhanced density, uniformity, and closer alignment to the ground truth LiDAR point clouds, which is the first of its kind network to do so. • Novel radar point cloud upsampling method using adversarial cascade approach. • Modification of CycleGAN for handling point clouds. • Dynamic batching to handle varying point clouds. • Enhanced radar-based point cloud method for improved people detection. [ABSTRACT FROM AUTHOR]

Universities worldwide are experiencing a surge in enrollments, therefore campus estate managers are seeking continuous data on attendance patterns to optimize the usage of classroom space. As a result, there is an increasing trend to measure classroom attendance by employing various sensing technologies, among which pervasive WiFi infrastructure is seen as a low-cost method. In a dense campus environment, the number of connected WiFi users does not well estimate room occupancy since connection counts are polluted by adjoining rooms, outdoor walkways, and network load balancing. This paper develops machine learning-based models, including unsupervised clustering and a combination of classification and regression algorithms, to infer classroom occupancy from WiFi sensing infrastructure. Our contributions are three-fold: (1) We analyze metadata from a dense and dynamic wireless network comprising of thousands of access points (APs) to draw insights into coverage of APs, the behavior of WiFi-connected users, and challenges of estimating room occupancy; (2) We propose a method to automatically map APs to classrooms and evaluate K-means, Expectation-Maximization (EM-GMM) and Hierarchical Clustering (HC) algorithms; and (3) We model classroom occupancy and evaluate varying algorithms, namely Logistic Regression, Support Vector Machine (SVM), Linear Discriminant Analysis (LDA), Linear Regression (LR) and Support Vector Regression (SVR). We achieve 84.6% accuracy in mapping APs to classrooms, while our estimation for room occupancy (with symmetric Mean Absolute Percentage Error (sMAPE) of 13.10%) is comparable to beam counter sensors. [ABSTRACT FROM AUTHOR]

Estimation of the presence of people in real time is extremely useful for businesses in providing better services. Many companies and researchers have attempted various researches in order to count the number of floating population in a specific space. Recently, as part of smart cities and digital twins, commercialization of measuring floating populations using Wi-Fi signals has become active in the public and private sectors. In this paper we present a method of estimating the floating population based on MAC-address values collected from smartphones. By distinguishing Real MAC-address and Random MAC-address values, we compare the estimated number of smartphone devices and the actual number of people caught on CCTV screens to evaluate the accuracy of the proposed method. And it appeared to have a similar correlation between the two datas. As a result, we present a method of estimating the floating population based on analyzing Wi-Fi Probe Requests. [ABSTRACT FROM AUTHOR]

Crowd monitoring was an essential measure to deal with over-tourism problems in urban destinations in the pre-COVID era. It will play a crucial role in the pandemic scenario when restarting tourism and making destinations safer. Notably, a Destination Management Organisation (DMO) of a smart destination needs to deploy a technological layer for crowd monitoring that allows data gathering in order to count visitors and distinguish them from residents. The correct identification of visitors versus residents by a DMO, while privacy rights (e.g., Regulation EU 2016/679, also known as GDPR) are ensured, is an ongoing problem that has not been fully solved. In this paper, we describe a novel approach to gathering crowd data by processing (i) massive scanning of WiFi access points of the smart destination to find SSIDs (Service Set Identifier), as well as (ii) the exposed Preferred Network List (PNL) containing the SSIDs of WiFi access points to which WiFi-enabled mobile devices are likely to connect. These data enable us to provide the number of visitors and residents of a crowd at a given point of interest of a tourism destination. A pilot study has been conducted in the city of Alcoi (Spain), comparing data from our approach with data provided by manually filled surveys from the Alcoi Tourist Info office, with an average accuracy of 83%, thus showing the feasibility of our policy to enrich the information system of a smart destination. [ABSTRACT FROM AUTHOR]

With the explosive development of artificial intelligence technology, machine learning, deep learning and other technologies have been widely used in face recognition, pedestrian detection, video tracking and other fields. Among them, using target detection for indoor crowd statistics has attracted a lot of attentions. Due to the problems such as mutual occlusion of crowds and blurred target features in the indoor monitoring screen, it often leads to low detection accuracy and high false detection rate and missed detection rate. In order to solve this problem, an indoor people counting model based on global attention is proposed. The model introduces the attention mechanism, optimizes the object detection algorithm YOLOv3, and enhances the detection ability by extracting more features of small or unclear heads. The experimental results show that the improved network model has higher recall and average precision. [ABSTRACT FROM AUTHOR]