Robust People Detection and Tracking from an Overhead Time-of-Flight Camera

Fernandez-Rincon, Alvaro; Fuentes-Jiménez, David; Losada, Cristina; Marrón, M.; Luna, Carlos; Macías-Guarasa, Javier; Mazo, Manuel

doi:10.5220/0006169905560564

Cited by 8 publications

(5 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• SLICE SV M [15]: this method is similar to SLICE P CA , but uses a SVM classifier instead of PCA. In our experiments, the SVM model is trained again using the full GOTPD1 dataset.…”

Section: Methodsmentioning

confidence: 99%

DPDnet: A robust people detector using deep learning with an overhead depth camera

Fuentes-Jimenez

Martín-López

Losada

et al. 2020

Expert Systems with Applications

Self Cite

View full text Add to dashboard Cite

In this paper we propose a method based on deep learning that detects multiple people from a single overhead depth image with high reliability. Our neural network, called DPDnet, is based on two fully-convolutional encoder-decoder neural blocks based on residual layers. The main block takes a depth image as input and generates a pixel-wise confidence map, where each detected person in the image is represented by a Gaussian-like distribution. The refinement block combines the depth image and the output from the main block, to refine the confidence map. Both blocks are simultaneously trained end-to-end using depth images and head position labels.The experimental work shows that DPDnet outperforms state-of-the-art methods, with accuracies greater than 99% in three different publicly available datasets, without retraining not fine-tuning. In addition, the computational complexity of our proposal is independent of the number of people in the scene and runs in real time using conventional GPUs.

show abstract

“…• SLICE SV M [15]: this method is similar to SLICE P CA , but uses a SVM classifier instead of PCA. In our experiments, the SVM model is trained again using the full GOTPD1 dataset.…”

Section: Methodsmentioning

confidence: 99%

DPDnet: A robust people detector using deep learning with an overhead depth camera

Fuentes-Jimenez

Martín-López

Losada

et al. 2020

Expert Systems with Applications

Self Cite

View full text Add to dashboard Cite

show abstract

“…This method can estimate both the cumulative count (i.e., the total count of people since the beginning of the video) and the instantaneous count (i.e., the count at any given time). The second method is called region of interest (ROI), which can estimate crowd density by evaluating the number of people who are present within a specific region of interest in the monitored scene [ 41 , 42 , 43 ].…”

Section: Background and Related Workmentioning

confidence: 99%

Edge-Based Transfer Learning for Classroom Occupancy Detection in a Smart Campus Context

Monti

Tse

Tang

et al. 2022

Sensors

View full text Add to dashboard Cite

Studies and systems that are aimed at the identification of the presence of people within an indoor environment and the monitoring of their activities and flows have been receiving more attention in recent years, specifically since the beginning of the COVID-19 pandemic. This paper proposes an approach for people counting that is based on the use of cameras and Raspberry Pi platforms, together with an edge-based transfer learning framework that is enriched with specific image processing strategies, with the aim of this approach being adopted in different indoor environments without the need for tailored training phases. The system was deployed on a university campus, which was chosen as the case study. The proposed system was able to work in classrooms with different characteristics. This paper reports a proposed architecture that could make the system scalable and privacy compliant and the evaluation tests that were conducted in different types of classrooms, which demonstrate the feasibility of this approach. Overall, the system was able to count the number of people in classrooms with a maximum mean absolute error of 1.23.

show abstract

“…A technique that could be adopted is based on the recognition of video segments stream, where the counting process can be divided into two main steps: (i) detecting moving blob on the basis of algorithms for motion detection (background subtraction and/or a segmentation strategy, i.e., K-means); (ii) monitoring the detected blobs, with the aim of identifying the direction of the monitor or to compute the number of people that are present in case of a single frame shot from a camera. In order to accomplish this task, two main categories of methods can be adopted: (i) the Line of Interest (LOI) counting methods can evaluate the number of people crossing a virtual Line of Interest within the monitored scene [32]; (ii) the Region of Interest (ROI) counting methods can estimate crowd, evaluating the number of people who are present within a specific Region of Interest in the monitored scene [33].…”

Section: Iot Infrastructures For Classroom Occupancy Monitoringmentioning

confidence: 99%

On exploiting Data Visualization and IoT for Increasing Sustainability and Safety in a Smart Campus

et al. 2021

View full text Add to dashboard Cite

In a world that is getting increasingly digital and interconnected, and where more and more physical objects are integrated into the information network (Internet of Things, IoT), Data Visualization can facilitate the understanding of huge volumes of data. In this paper, we present the design and implementation of a testbed where IoT and Data Visualization have been exploited to increase the sustainability and safety of the Cesena (Smart) Campus. In particular, we detail the overall system architecture and the interactive dashboard that facilitates the management of the campus premises and the timetabling. Exploiting our system, we show how we can improve the campus sustainability (in terms of energy saving) and safety (considering the COVID-19 restrictions and regulations).

show abstract

Robust People Detection and Tracking from an Overhead Time-of-Flight Camera

Cited by 8 publications

References 14 publications

DPDnet: A robust people detector using deep learning with an overhead depth camera

DPDnet: A robust people detector using deep learning with an overhead depth camera

Edge-Based Transfer Learning for Classroom Occupancy Detection in a Smart Campus Context

On exploiting Data Visualization and IoT for Increasing Sustainability and Safety in a Smart Campus

Contact Info

Product

Resources

About