Making the Invisible Visible: Action Recognition Through Walls and Occlusions
Understanding people's actions and interactions typically depends on seeing them. Automating the process of action recognition from visual data has been the topic of much research in the computer vision community. But what if it is too dark, or if the person is occluded or behind a wall? In thi...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers (IEEE)
2021
|
| Online Access: | https://hdl.handle.net/1721.1/129445 |
| _version_ | 1811092198345670656 |
|---|---|
| author | Li, Tianhong Fan, Lijie Zhao, Mingmin Liu, Yingcheng Katabi, Dina |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Li, Tianhong Fan, Lijie Zhao, Mingmin Liu, Yingcheng Katabi, Dina |
| author_sort | Li, Tianhong |
| collection | MIT |
| description | Understanding people's actions and interactions typically depends on seeing them. Automating the process of action recognition from visual data has been the topic of much research in the computer vision community. But what if it is too dark, or if the person is occluded or behind a wall? In this paper, we introduce a neural network model that can detect human actions through walls and occlusions, and in poor lighting conditions. Our model takes radio frequency (RF) signals as input, generates 3D human skeletons as an intermediate representation, and recognizes actions and interactions of multiple people over time. By translating the input to an intermediate skeleton-based representation, our model can learn from both vision-based and RF-based datasets, and allow the two tasks to help each other. We show that our model achieves comparable accuracy to vision-based action recognition systems in visible scenarios, yet continues to work accurately when people are not visible, hence addressing scenarios that are beyond the limit of today's vision-based action recognition. |
| first_indexed | 2024-09-23T15:14:32Z |
| format | Article |
| id | mit-1721.1/129445 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T15:14:32Z |
| publishDate | 2021 |
| publisher | Institute of Electrical and Electronics Engineers (IEEE) |
| record_format | dspace |
| spelling | mit-1721.1/1294452022-10-02T01:37:17Z Making the Invisible Visible: Action Recognition Through Walls and Occlusions Li, Tianhong Fan, Lijie Zhao, Mingmin Liu, Yingcheng Katabi, Dina Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Understanding people's actions and interactions typically depends on seeing them. Automating the process of action recognition from visual data has been the topic of much research in the computer vision community. But what if it is too dark, or if the person is occluded or behind a wall? In this paper, we introduce a neural network model that can detect human actions through walls and occlusions, and in poor lighting conditions. Our model takes radio frequency (RF) signals as input, generates 3D human skeletons as an intermediate representation, and recognizes actions and interactions of multiple people over time. By translating the input to an intermediate skeleton-based representation, our model can learn from both vision-based and RF-based datasets, and allow the two tasks to help each other. We show that our model achieves comparable accuracy to vision-based action recognition systems in visible scenarios, yet continues to work accurately when people are not visible, hence addressing scenarios that are beyond the limit of today's vision-based action recognition. 2021-01-19T16:50:44Z 2021-01-19T16:50:44Z 2020-02 2019-10 2020-12-23T16:17:18Z Article http://purl.org/eprint/type/ConferencePaper 9781728148038 https://hdl.handle.net/1721.1/129445 Li, Tianhong et al. "Making the Invisible Visible: Action Recognition Through Walls and Occlusions." 2019 IEEE/CVF International Conference on Computer Vision (ICCV), October-November 2019, Seoul, Korea, Institute of Electrical and Electronics Engineers, February 2020. © 2019 IEEE en http://dx.doi.org/10.1109/iccv.2019.00096 2019 IEEE/CVF International Conference on Computer Vision (ICCV) Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Institute of Electrical and Electronics Engineers (IEEE) arXiv |
| spellingShingle | Li, Tianhong Fan, Lijie Zhao, Mingmin Liu, Yingcheng Katabi, Dina Making the Invisible Visible: Action Recognition Through Walls and Occlusions |
| title | Making the Invisible Visible: Action Recognition Through Walls and Occlusions |
| title_full | Making the Invisible Visible: Action Recognition Through Walls and Occlusions |
| title_fullStr | Making the Invisible Visible: Action Recognition Through Walls and Occlusions |
| title_full_unstemmed | Making the Invisible Visible: Action Recognition Through Walls and Occlusions |
| title_short | Making the Invisible Visible: Action Recognition Through Walls and Occlusions |
| title_sort | making the invisible visible action recognition through walls and occlusions |
| url | https://hdl.handle.net/1721.1/129445 |
| work_keys_str_mv | AT litianhong makingtheinvisiblevisibleactionrecognitionthroughwallsandocclusions AT fanlijie makingtheinvisiblevisibleactionrecognitionthroughwallsandocclusions AT zhaomingmin makingtheinvisiblevisibleactionrecognitionthroughwallsandocclusions AT liuyingcheng makingtheinvisiblevisibleactionrecognitionthroughwallsandocclusions AT katabidina makingtheinvisiblevisibleactionrecognitionthroughwallsandocclusions |