Optimized spatio-temporal descriptors for real-time fall detection: comparison of support vector machine and Adaboost-based classification

Charfi, Imen; Mitéran, Johel; Dubois, Julien; Atri, Mohamed; Tourki, Rached

doi:10.1117/1.jei.22.4.041106

Cited by 141 publications

(89 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We recall in the next section the main characteristics of the method which has been described and evaluated in detail in [9].…”

Section: Fall Detectionmentioning

confidence: 99%

“…This approach is improving state-of-the-art fall detection performance using a single-camera system [9]. Therefore, it has been selected as a candidate for implementation.…”

Section: Fall Detection Performancementioning

confidence: 99%

“…Performance of our fall detection method and the used metric are fully described in [9], including several protocols allowing for example (i) to study the robustness of the detection against location change between training and testing, and (ii) to illustrate how the non-fall activities are confused with fall. We evaluate the fall detection process after the filtering, taking into account a slot of w ¼ 18 images, computing the sensitivity, the specificity, the accuracy, the recall and the final classification error rate as follows: …”

Section: Fall Detection Performancementioning

confidence: 99%

“…We studied four transformations : the discrete first derivative which captures the velocity, the discrete second derivative to capture acceleration, and the windowed Fourier Transform or wavelets coefficients. After feature selection, only the X signal and the first derivative were retained [9]. The result of the first derivative operator noted FD applied to X (set of u signals) during w 1 frames is FDðXÞ a set of w 1 Â u values.…”

Section: Low-level Feature Transformation and Selectionmentioning

confidence: 99%

See 3 more Smart Citations

Fast prototyping of a SoC-based smart-camera: a real-time fall detection case study

Senouci

Charfi

Heyrman

et al. 2014

J Real-Time Image Proc

Self Cite

View full text Add to dashboard Cite

Smart camera, i.e. cameras that are able to acquire and process images in real-time, is a typical example of the new embedded computer vision systems. A key example of application is automatic fall detection, which can be useful for helping elderly people in daily life. In this paper, we propose a methodology for development and fast-prototyping of a fall detection system based on such a smart camera, which allows to reduce the development time compared to standard approaches. Founded on a supervised classification approach, we propose a HW/SW implementation to detect falls in a home environment using a single camera and an optimized descriptor adapted to real-time tasks. This heterogeneous implementation is based on Xilinx's system-on-chip named Zynq. The main contributions of this work are (i) the proposal of a codesign methodology. These methodologies enable the HW/ SW partitioning to be delayed using high-level algorithmic description and high-level synthesis tools. Our approach enables fast prototyping which allows fast architecture exploration and optimisation to be performed, (ii) the design of a hardware accelerator dedicated to boostingbased classification, which is a very popular and efficient algorithm used in image analysis, (iii) the proposal of falldetection embedded in a smart camera and enabling integration into the elderly people environment. Performances of our system are finally compared to the state-of-the-art.

show abstract

“…We recall in the next section the main characteristics of the method which has been described and evaluated in detail in [9].…”

Section: Fall Detectionmentioning

confidence: 99%

“…This approach is improving state-of-the-art fall detection performance using a single-camera system [9]. Therefore, it has been selected as a candidate for implementation.…”

Section: Fall Detection Performancementioning

confidence: 99%

Section: Fall Detection Performancementioning

confidence: 99%

Section: Low-level Feature Transformation and Selectionmentioning

confidence: 99%

See 2 more Smart Citations

Fast prototyping of a SoC-based smart-camera: a real-time fall detection case study

Senouci

Charfi

Heyrman

et al. 2014

J Real-Time Image Proc

Self Cite

View full text Add to dashboard Cite

show abstract

“…Charfi et al [18,19] introduced an optimized spatio-temporal human fall descriptor, named STHF, which uses several combinations of transformations of geometrical features. The extracted features, such as, width and height of human body bounding box, projection histograms and the user's trajectory, then were used for supervised SVM and Adaboost classifiers.…”

Section: Vision-based Devisesmentioning

confidence: 99%

A Panoramic Study of Fall Detection Technologies

2016

IJCSI

View full text Add to dashboard Cite

Falls are a major risk of injury for elderly aged 65 or over, blind people, people with balance disorder and leg weakness. In this regard, assistive technology which aims to identify fall events at real time can reduce the rate of impairments and mortality. This study offer a literature research reference value for bioengineers for further research. Much of the past and the current fall detection research, the vital signals features and the way features are extracted and fed to a classifier are introduced. The study concludes with an assessment of the current technologies highlighting their critical limitations along with suggestions for future research direction in this rapidly developing field of study.

show abstract

An Empirical Study Towards Understanding How Deep Convolutional Nets Recognize Falls

Zhang

Neumann

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Detecting unintended falls is essential for ambient intelligence and healthcare of elderly people living alone. In recent years, deep convolutional nets are widely used in human action analysis, based on which a number of fall detection methods have been proposed. Despite their highly effective performances, the behaviors of how the convolutional nets recognize falls are still not clear. In this paper, instead of proposing a novel approach, we perform a systematical empirical study, attempting to investigate the underlying fall recognition process. We propose four tasks to investigate, which involve five types of input modalities, seven net instances and different training samples. The obtained quantitative and qualitative results reveal the patterns that the nets tend to learn, and several factors that can heavily influence the performances on fall recognition. We expect that our conclusions are favorable to proposing better deep learning solutions to fall detection systems.

show abstract

Optimized spatio-temporal descriptors for real-time fall detection: comparison of support vector machine and Adaboost-based classification

Cited by 141 publications

References 16 publications

Fast prototyping of a SoC-based smart-camera: a real-time fall detection case study

Fast prototyping of a SoC-based smart-camera: a real-time fall detection case study

A Panoramic Study of Fall Detection Technologies

An Empirical Study Towards Understanding How Deep Convolutional Nets Recognize Falls

Contact Info

Product

Resources

About