Assessing the Quality of Actions
While recent advances in computer vision have provided reliable methods to recognize actions in both images and videos, the problem of assessing how well people perform actions has been largely unexplored in computer vision. Since methods for assessing action quality have many real-world application...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | en_US |
| Published: |
Springer-Verlag
2014
|
| Online Access: | http://hdl.handle.net/1721.1/90990 https://orcid.org/0000-0003-4915-0256 https://orcid.org/0000-0001-5676-2387 |
| _version_ | 1811071356340535296 |
|---|---|
| author | Pirsiavash, Hamed Torralba, Antonio Vondrick, Carl Martin |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Pirsiavash, Hamed Torralba, Antonio Vondrick, Carl Martin |
| author_sort | Pirsiavash, Hamed |
| collection | MIT |
| description | While recent advances in computer vision have provided reliable methods to recognize actions in both images and videos, the problem of assessing how well people perform actions has been largely unexplored in computer vision. Since methods for assessing action quality have many real-world applications in healthcare, sports, and video retrieval, we believe the computer vision community should begin to tackle this challenging problem. To spur progress, we introduce a learning-based framework that takes steps towards assessing how well people perform actions in videos. Our approach works by training a regression model from spatiotemporal pose features to scores obtained from expert judges. Moreover, our approach can provide interpretable feedback on how people can improve their action. We evaluate our method on a new Olympic sports dataset, and our experiments suggest our framework is able to rank the athletes more accurately than a non-expert human. While promising, our method is still a long way to rivaling the performance of expert judges, indicating that there is significant opportunity in computer vision research to improve on this difficult yet important task. |
| first_indexed | 2024-09-23T08:49:55Z |
| format | Article |
| id | mit-1721.1/90990 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T08:49:55Z |
| publishDate | 2014 |
| publisher | Springer-Verlag |
| record_format | dspace |
| spelling | mit-1721.1/909902022-09-23T14:49:53Z Assessing the Quality of Actions Pirsiavash, Hamed Torralba, Antonio Vondrick, Carl Martin Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Pirsiavash, Hamed Vondrick, Carl Martin Torralba, Antonio While recent advances in computer vision have provided reliable methods to recognize actions in both images and videos, the problem of assessing how well people perform actions has been largely unexplored in computer vision. Since methods for assessing action quality have many real-world applications in healthcare, sports, and video retrieval, we believe the computer vision community should begin to tackle this challenging problem. To spur progress, we introduce a learning-based framework that takes steps towards assessing how well people perform actions in videos. Our approach works by training a regression model from spatiotemporal pose features to scores obtained from expert judges. Moreover, our approach can provide interpretable feedback on how people can improve their action. We evaluate our method on a new Olympic sports dataset, and our experiments suggest our framework is able to rank the athletes more accurately than a non-expert human. While promising, our method is still a long way to rivaling the performance of expert judges, indicating that there is significant opportunity in computer vision research to improve on this difficult yet important task. National Science Foundation (U.S.). Graduate Research Fellowship Google (Firm) (Research Award) United States. Office of Naval Research. Multidisciplinary University Research Initiative (N000141010933) 2014-10-20T16:48:38Z 2014-10-20T16:48:38Z 2014 Article http://purl.org/eprint/type/ConferencePaper 978-3-319-10598-7 978-3-319-10599-4 0302-9743 1611-3349 http://hdl.handle.net/1721.1/90990 Pirsiavash, Hamed, Carl Vondrick, and Antonio Torralba. “Assessing the Quality of Actions.” Lecture Notes in Computer Science (2014): 556–571. https://orcid.org/0000-0003-4915-0256 https://orcid.org/0000-0001-5676-2387 en_US http://dx.doi.org/10.1007/978-3-319-10599-4_36 Computer Vision – ECCV 2014 Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Springer-Verlag MIT web domain |
| spellingShingle | Pirsiavash, Hamed Torralba, Antonio Vondrick, Carl Martin Assessing the Quality of Actions |
| title | Assessing the Quality of Actions |
| title_full | Assessing the Quality of Actions |
| title_fullStr | Assessing the Quality of Actions |
| title_full_unstemmed | Assessing the Quality of Actions |
| title_short | Assessing the Quality of Actions |
| title_sort | assessing the quality of actions |
| url | http://hdl.handle.net/1721.1/90990 https://orcid.org/0000-0003-4915-0256 https://orcid.org/0000-0001-5676-2387 |
| work_keys_str_mv | AT pirsiavashhamed assessingthequalityofactions AT torralbaantonio assessingthequalityofactions AT vondrickcarlmartin assessingthequalityofactions |