S3D-CNN: skeleton-based 3D consecutive-low-pooling neural network for fall detection

Xiong, Xin; Min, Weidong; Zheng, Wei‐Shi; Liao, Pin; Yang, Hao; Wang, Shuai

doi:10.1007/s10489-020-01751-y

Cited by 34 publications

(9 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A single video is typically divided into one to three shot parts given the characteristics of datasets. We then reference the method in [ 47 ] to compare the entropy of different shot parts. The video shot with maximum entropy contains nearly complete action information.…”

Section: Action Sequences Optimization Methodsmentioning

confidence: 99%

Action Recognition Using Action Sequences Optimization and Two-Stream 3D Dilated Neural Network

Xiong

Min

Han

et al. 2022

Computational Intelligence and Neuroscience

Self Cite

View full text Add to dashboard Cite

Effective extraction and representation of action information are critical in action recognition. The majority of existing methods fail to recognize actions accurately because of interference of background changes when the proportion of high-activity action areas is not reinforced and by using RGB flow alone or combined with optical flow. A novel recognition method using action sequences optimization and two-stream fusion network with different modalities is proposed to solve these problems. The method is based on shot segmentation and dynamic weighted sampling, and it reconstructs the video by reinforcing the proportion of high-activity action areas, eliminating redundant intervals, and extracting long-range temporal information. A two-stream 3D dilated neural network that integrates features of RGB and human skeleton information is also proposed. The human skeleton information strengthens the deep representation of humans for robust processing, alleviating the interference of background changes, and the dilated CNN enlarges the receptive field of feature extraction. Compared with existing approaches, the proposed method achieves superior or comparable classification accuracies on benchmark datasets UCF101 and HMDB51.

show abstract

Section: Action Sequences Optimization Methodsmentioning

confidence: 99%

Action Recognition Using Action Sequences Optimization and Two-Stream 3D Dilated Neural Network

Xiong

Min

Han

et al. 2022

Computational Intelligence and Neuroscience

Self Cite

View full text Add to dashboard Cite

show abstract

“…A two-level probability model is used to generate better skeleton sequences, and then RNN network is used for sign language recognition. Xiong et al [35] proposed a skeleton-based 3D consecutive-low-pooling neural network (S3D-CNN) for fall detection.An activity feature clustering selector and a 3Dconsecutive-low-pooling (3D-CLP) neural network are designed to recognition for fall detection. The model framework proposed by Nunez et al [22] is composed of convolutional neural network and long-term and short-term memory recurrent network.…”

Section: Proposed a Skeleton-based Csl Recognition And Generation Framework Based On Recurrent Neural Network (Rnn)mentioning

confidence: 99%

Sign language recognition based on skeleton with K3D-LSTMresidual networks

Han

Huangfu

Min

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Most existing deep learning-based dynamic sign language recognition methods directly use either the video sequence based on RGB information, or whole sequences instead of only the video sequence that represents the change of gesture. These characteristics lead to inaccurate extraction of hand gesture features and failure to achieve good recognition accuracy for complex gestures. In order to solve these problems, this paper proposes a new method of dynamic hand gesture recognition for key skeleton information, which combines residual convolutional neural network and long short-term memory recurrent network, which is called KLSTM-3D residual network (K3D ResNet). In K3DResNet, the spatiotemporal complexity of network computation is reduced by extracting the representative skeleton frame of gesture change. Then, the spatiotemporal features are extracted from the skeleton keyframe sequence, and the intermediate score corresponding to each action in the video sequence is established after the feature analysis. Finally, the classification of video sequences can accurately identify sign language. Experiments were performed on datasets DHG14/28 and SHREC’17 Track. The accuracy of verification on dataset DEVISIGN D reached 88.6%. In addition, the accuracy of the combination of RGB and skeleton information reached 93.2%.

show abstract

“…However, the performance of those classifiers reduces a complicated situation, such as the same color of clothes and background, multisource, and multiple people with occlusion. Xiong 8 proposed a skeleton-based three-dimensional consecutive-low-pooling neural network (S3D-CNN) for fall detection. The proposed system was evaluated on public and self-collected datasets and achieved best results compared to existing methods.…”

Section: Fall Detectionmentioning

confidence: 99%

Deep learning for vision‐based fall detection system: Enhanced optical dynamic flow

Chhetri

Alsadoon

et al. 2020

Computational Intelligence

View full text Add to dashboard Cite

Accurate fall detection for the assistance of older people is crucial to reduce incidents of deaths or injuries due to falls. Meanwhile, vision-based fall detection system has shown some significant results to detect falls. Still, numerous challenges need to be resolved. The impact of deep learning has changed the landscape of the vision-based system, such as action recognition. The deep learning technique has not been successfully implemented in vision-based fall detection system due to the requirement of a large amount of computation power and requirement of a large amount of sample training data. This research aims to propose a vision-based fall detection system that improves the accuracy of fall detection in some complex environments such as the change of light condition in the room. Also, this research aims to increase the performance of the pre-processing of video images. The proposed system consists of Enhanced Dynamic Optical Flow technique that encodes the temporal data of optical flow videos by the method of rank pooling, which thereby improves the processing time of fall detection and improves the classification accuracy in dynamic lighting condition. The experimental results showed that the classification accuracy of the fall detection improved by around 3% and the processing time by 40-50 ms. The proposed system concentrates on decreasing the processing time 578

show abstract

S3D-CNN: skeleton-based 3D consecutive-low-pooling neural network for fall detection

Cited by 34 publications

References 52 publications

Action Recognition Using Action Sequences Optimization and Two-Stream 3D Dilated Neural Network

Action Recognition Using Action Sequences Optimization and Two-Stream 3D Dilated Neural Network

Sign language recognition based on skeleton with K3D-LSTMresidual networks

Deep learning for vision‐based fall detection system: Enhanced optical dynamic flow

Contact Info

Product

Resources

About