Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/151813
Title: | Anàlisi d’emocions multitasca |
Author: | Ostos Balfagón, Rubén |
Tutor: | Torres-Sospedra, Joaquín ![]() Benito Altamirano, Ismael ![]() |
Others: | Acedo Nadal, Susana |
Abstract: | Emotion analysis and detection is one of the most widespread fields of computer vision, but also of speech and text analysis. It has multiple fields of application, and it is a field where different sciences intersect and try to join forces in order to obtain valid applications, given its multiple applications, be it for medicine, fatigue detection, monitoring, online learning, lie detection or for legal applications, etc. There are different approaches to deal with emotion detection, the most common being the recognition of basic emotions, the estimation of continuous affect and the detection of facial units. One of the approaches in recent years has been to use combinations of these in order to improve the results, using multi-task techniques. The purpose of this work is to create one or more multi-task emotion detection models. Throughout this work, we will be able to see how the state of the art in the matter is evaluated and, using a case of success as a basis, improvements of different types are applied in order to try to obtain a model with a superior performance. The improvements will be based on the data available in the different state of the art papers and an exhaustive study of the results of these will be carried out. At the end of this work, we will find a model able to compete with the best participants of the 7th ABAW as one of the experiments will provide an improved version of the model used as the basis. |
Keywords: | emotion detection deep learning neural networks |
Document type: | info:eu-repo/semantics/masterThesis |
Issue Date: | Dec-2024 |
Publication license: | http://creativecommons.org/licenses/by-nc-nd/3.0/es/ ![]() |
Appears in Collections: | Bachelor thesis, research projects, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
rostosbTFM1224.pdf | 2,69 MB | Adobe PDF | ![]() View/Open |
Share:


This item is licensed under aCreative Commons License