Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall
Unlike most of the existing neural network-based fall detection methods, which only detect fall at the time range, the algorithm proposed in this paper detect fall in both spatial and temporal dimension. A movement tube detection network integrating 3D CNN and object detection framework such as SSD...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-04-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/10/8/898 |
| _version_ | 1797538273103446016 |
|---|---|
| author | Song Zou Weidong Min Lingfeng Liu Qi Wang Xiang Zhou |
| author_facet | Song Zou Weidong Min Lingfeng Liu Qi Wang Xiang Zhou |
| author_sort | Song Zou |
| collection | DOAJ |
| description | Unlike most of the existing neural network-based fall detection methods, which only detect fall at the time range, the algorithm proposed in this paper detect fall in both spatial and temporal dimension. A movement tube detection network integrating 3D CNN and object detection framework such as SSD is proposed to detect human fall with constrained movement tubes. The constrained movement tube, which encapsulates the person with a sequence of bounding boxes, has the merits of encapsulating the person closely and avoiding peripheral interference. A 3D convolutional neural network is used to encode the motion and appearance features of a video clip, which are fed into the tube anchors generation layer, softmax classification, and movement tube regression layer. The movement tube regression layer fine tunes the tube anchors to the constrained movement tubes. A large-scale spatio-temporal (LSST) fall dataset is constructed using self-collected data to evaluate the fall detection in both spatial and temporal dimensions. LSST has three characteristics of large scale, annotation, and posture and viewpoint diversities. Furthermore, the comparative experiments on a public dataset demonstrate that the proposed algorithm achieved sensitivity, specificity an accuracy of 100%, 97.04%, and 97.23%, respectively, outperforms the existing methods. |
| first_indexed | 2024-03-10T12:29:11Z |
| format | Article |
| id | doaj.art-a0891c0cdec64d649605c6ed985caa7d |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T12:29:11Z |
| publishDate | 2021-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-a0891c0cdec64d649605c6ed985caa7d2023-11-21T14:49:45ZengMDPI AGElectronics2079-92922021-04-0110889810.3390/electronics10080898Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect FallSong Zou0Weidong Min1Lingfeng Liu2Qi Wang3Xiang Zhou4School of Information Engineering, Nanchang University, Nanchang 330031, ChinaSchool of Software, Nanchang University, Nanchang 330047, ChinaSchool of Software, Nanchang University, Nanchang 330047, ChinaSchool of Information Engineering, Nanchang University, Nanchang 330031, ChinaSchool of Software, Nanchang University, Nanchang 330047, ChinaUnlike most of the existing neural network-based fall detection methods, which only detect fall at the time range, the algorithm proposed in this paper detect fall in both spatial and temporal dimension. A movement tube detection network integrating 3D CNN and object detection framework such as SSD is proposed to detect human fall with constrained movement tubes. The constrained movement tube, which encapsulates the person with a sequence of bounding boxes, has the merits of encapsulating the person closely and avoiding peripheral interference. A 3D convolutional neural network is used to encode the motion and appearance features of a video clip, which are fed into the tube anchors generation layer, softmax classification, and movement tube regression layer. The movement tube regression layer fine tunes the tube anchors to the constrained movement tubes. A large-scale spatio-temporal (LSST) fall dataset is constructed using self-collected data to evaluate the fall detection in both spatial and temporal dimensions. LSST has three characteristics of large scale, annotation, and posture and viewpoint diversities. Furthermore, the comparative experiments on a public dataset demonstrate that the proposed algorithm achieved sensitivity, specificity an accuracy of 100%, 97.04%, and 97.23%, respectively, outperforms the existing methods.https://www.mdpi.com/2079-9292/10/8/898human fall detectionmovement tube detection networkspatial and temporal dimensionsself-collected large-scale fall detection dataset |
| spellingShingle | Song Zou Weidong Min Lingfeng Liu Qi Wang Xiang Zhou Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall Electronics human fall detection movement tube detection network spatial and temporal dimensions self-collected large-scale fall detection dataset |
| title | Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall |
| title_full | Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall |
| title_fullStr | Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall |
| title_full_unstemmed | Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall |
| title_short | Movement Tube Detection Network Integrating 3D CNN and Object Detection Framework to Detect Fall |
| title_sort | movement tube detection network integrating 3d cnn and object detection framework to detect fall |
| topic | human fall detection movement tube detection network spatial and temporal dimensions self-collected large-scale fall detection dataset |
| url | https://www.mdpi.com/2079-9292/10/8/898 |
| work_keys_str_mv | AT songzou movementtubedetectionnetworkintegrating3dcnnandobjectdetectionframeworktodetectfall AT weidongmin movementtubedetectionnetworkintegrating3dcnnandobjectdetectionframeworktodetectfall AT lingfengliu movementtubedetectionnetworkintegrating3dcnnandobjectdetectionframeworktodetectfall AT qiwang movementtubedetectionnetworkintegrating3dcnnandobjectdetectionframeworktodetectfall AT xiangzhou movementtubedetectionnetworkintegrating3dcnnandobjectdetectionframeworktodetectfall |