| dc.creator | Straminsky, Axel | |
| dc.creator | Jacobo, Julio | |
| dc.creator | Buemi, María Elena | |
| dc.date | 2021-10 | |
| dc.date | 2021 | |
| dc.date | 2022-08-30T16:59:31Z | |
| dc.date.accessioned | 2023-07-15T07:47:43Z | |
| dc.date.available | 2023-07-15T07:47:43Z | |
| dc.identifier | http://sedici.unlp.edu.ar/handle/10915/141291 | |
| dc.identifier | http://50jaiio.sadio.org.ar/pdfs/saiv/SAIV-08.pdf | |
| dc.identifier | issn:2683-8990 | |
| dc.identifier.uri | https://repositorioslatinoamericanos.uchile.cl/handle/2250/7481862 | |
| dc.description | The goal of this work is to propose possible improvements on one of the latest models for Video Action Recognition based on currently existing attention mechanisms. We took a model architecture that uses 2 sub-models in paralell: one based on Optical Flow and the other based on the video itself, and proposed the following improvements: adding mixed precision in the training loop, using a Ranger optimizer instead of SGD, and expanding the Attention Mechanism. The video database used for this work was the EGTEA+ that is a action database of first person videos of daily activities. | |
| dc.description | Sociedad Argentina de Informática e Investigación Operativa | |
| dc.format | application/pdf | |
| dc.format | 36-39 | |
| dc.language | en | |
| dc.rights | http://creativecommons.org/licenses/by-nc-sa/3.0/ | |
| dc.rights | Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported (CC BY-NC-SA 3.0) | |
| dc.subject | Ciencias Informáticas | |
| dc.subject | First Person Vision | |
| dc.subject | Human Computer Interaction | |
| dc.subject | Action Recognition | |
| dc.subject | Attention module | |
| dc.title | An efficient action detection from first person vision with attention model | |
| dc.type | Objeto de conferencia | |
| dc.type | Objeto de conferencia | |
