Head Detection in Stereo Data for People Counting and Segmentation

Oosterhout, Tim van; Bakkes, Sander; Kröse, B.J.A.

doi:10.5220/0003362806200625

Cited by 10 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using images from depth sensing camera, the human tracking and detection systems have been presented by Wu et al [18] and Wetzel et al [19]. Hough circle was researched by Van Oosterhout, Bakkes, and Krose [20], while facts about the human head's and whorl shaped hairs have been studied by Nakatani et al [21]. Color information has been employed by Wateosot et al in [22], Nakatani et al in [21], and Gao et al in [23], while edge information has been used by others, such as Sobel filter or Canny edge detector [24].…”

Section: Related Workmentioning

confidence: 99%

Realtime Crowd Monitoring—Estimating Count, Speed and Direction of People Using Hybridized YOLOv4

Khel

Kadir

Khan³

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Researchers are becoming more interested in crowd surveillance because of its several potential applications. These applications may include detecting unusual activity for security purposes, monitoring reasons for archiving records, and conducting inventory for facility planning and extension. Detecting people and tracking them from a security viewpoint and understanding their behavior in places large crowds is highly important because unruly crowds in public spaces can lead to serious health and security concerns. Crowd related accidents happen to cause injuries and deaths, which often occur during events not properly planned. The planning of the organizers relies heavily on exploring the behavior of the few in a crowd of individuals and groups in thousands that create the crowds. It is this focus that provides the main reason for this research. This work proposes a model that can count people in crowds, automatically detect and track people, and then estimate their direction and speed. Deep learning networks have proven costly to run, needing memory and power to perform computations beyond what is possible on edge devices with limited resources. As a result, we propose the use of hybrid YOLOv4 consisting of detection method combined with the training phase pruning and the use the convolution attention module strategy. Accuracy of the Hybrid YOLOv4 is increased by 33%, whereas mAP reached 92.1%. While training on the JHU dataset, the suggested hybrid YOLOv4 strategy decreases the computational memory requirements, all of which closely meet the real-time application conditions. This work will help avoid the threatening situation of crowding gathering around to cause stampedes and thus risking crowds with disastrous consequences.

show abstract

Section: Related Workmentioning

confidence: 99%

Realtime Crowd Monitoring—Estimating Count, Speed and Direction of People Using Hybridized YOLOv4

Khel

Kadir

Khan³

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Data fusion of environmental sensors were confirmed to validate individual sensors for improved performance [7]. Some previous studies on estimating the number of people present in buildings utilized red, green, and blue (RGB) camera thermal arrays [8][9][10][11][12][13][14][15][16][17][18]. A head detection method with an accuracy ranging from 90 to 95% for different scenarios was proposed by Oosterhout and coauthors based on range data from stereo cameras for counting people [18].…”

Section: Related Workmentioning

confidence: 99%

“…Some previous studies on estimating the number of people present in buildings utilized red, green, and blue (RGB) camera thermal arrays [8][9][10][11][12][13][14][15][16][17][18]. A head detection method with an accuracy ranging from 90 to 95% for different scenarios was proposed by Oosterhout and coauthors based on range data from stereo cameras for counting people [18]. In contrast, we rely on thermal cameras to preserve people's privacy and to be able to capture the number of people present in thermal images even during the night.…”

Section: Related Workmentioning

confidence: 99%

Thermal Imagery Feature Extraction Techniques and the Effects on Machine Learning Models for Smart HVAC Efficiency in Building Energy

et al. 2021

View full text Add to dashboard Cite

The control of thermostats of a heating, ventilation, and air-conditioning (HVAC) system installed in commercial and residential buildings remains a pertinent problem in building energy efficiency and thermal comfort research. The ability to determine the number of people at a particular time in an area is imperative for energy efficiency in order to condition only occupied regions and thermally deficient regions. In this study of the best features comparison for detecting the number of people in an area, feature extraction techniques including wavelet scattering, wavelet decomposition, grey-level co-occurrence matrix (GLCM) and feature maps convolution neural network (CNN) layers were explored using thermal camera imagery. Specifically, the pretrained CNN networks explored are the deep residual (Resnet-50) and visual geometry group (VGG-16) networks. The discriminating potential of Haar, Daubechies and Symlets wavelet statistics on different distributions of data were investigated. The performance of VGG-16 and ResNet-50 in an end-to-end manner utilizing transfer learning approach was investigated. Experimental results showed the classification and regression trees (CART) model trained on only GLCM and Haar wavelet statistic features, individually achieved accuracies of approximately 80% and 84%, respectively, in the detection problem. Moreover, k-nearest neighbors (KNN) trained on the combined features of GLCM and Haar wavelet statistics achieved an accuracy of approximately 86%. In addition, the performance accuracy of the multi classification support vector machine (SVM) trained on deep features obtained from layers of pretrained ResNet-50 and VGG-16 was between 96% and 97%. Furthermore, ResNet-50 transfer learning outperformed the VGG-16 transfer learning model for occupancy detection using thermal imagery. Overall, the SVM model trained on features extracted from wavelet scattering emerged as the best performing classifier with an accuracy of 100%. A principal component analysis (PCA) on the wavelet scattering features proved that the first twenty (20) principal components achieved a similar accuracy level instead of training on the whole feature set to reduce the execution time. The occupancy detection models can be integrated into HVAC control systems for energy efficiency and security systems, and aid in the distribution of resources to people in an area.

show abstract

“…Lempitsky and Zisserman developed a supervised learning method able to count the number of people in surveillance video frames [8]. Oosterhout, Bakkes, and Kröse developed a head detection method capable of counting people from stereo camera data [9]. Wang and Jin proposed ways of estimating indoor occupancy by analyzing the carbon dioxide concentration of the return air of a room's ventilation system [10].…”

Section: Introductionmentioning

confidence: 99%

Synthetic Generation of Passive Infrared Motion Sensor Data Using a Game Engine

Stefansson

Karlsson

Persson

2021

Sensors

View full text Add to dashboard Cite

Quantifying the number of occupants in an indoor space is useful for a wide variety of applications. Attempts have been made at solving the task using passive infrared (PIR) motion sensor data together with supervised learning methods. Collecting a large labeled dataset containing both PIR motion sensor data and ground truth people count is however time-consuming, often requiring one hour of observation for each hour of data gathered. In this paper, a method is proposed for generating such data synthetically. A simulator is developed in the Unity game engine capable of producing synthetic PIR motion sensor data by detecting simulated occupants. The accuracy of the simulator is tested by replicating a real-world meeting room inside the simulator and conducting an experiment where a set of choreographed movements are performed in the simulated environment as well as the real room. In 34 out of 50 tested situations, the output from the simulated PIR sensors is comparable to the output from the real-world PIR sensors. The developed simulator is also used to study how a PIR sensor’s output changes depending on where in a room a motion is carried out. Through this, the relationship between sensor output and spatial position of a motion is discovered to be highly non-linear, which highlights some of the difficulties associated with mapping PIR data to occupancy count.

show abstract

Head Detection in Stereo Data for People Counting and Segmentation

Cited by 10 publications

References 20 publications

Realtime Crowd Monitoring—Estimating Count, Speed and Direction of People Using Hybridized YOLOv4

Realtime Crowd Monitoring—Estimating Count, Speed and Direction of People Using Hybridized YOLOv4

Thermal Imagery Feature Extraction Techniques and the Effects on Machine Learning Models for Smart HVAC Efficiency in Building Energy

Synthetic Generation of Passive Infrared Motion Sensor Data Using a Game Engine

Contact Info

Product

Resources

About