Por favor, use este identificador para citar o enlazar este ítem:
http://cimat.repositorioinstitucional.mx/jspui/handle/1008/1098| Models for a Quantitative Evaluation of Karate Motion Skills | |
| Roberto Dominguez Osuna | |
| Acceso Abierto | |
| Atribución-NoComercial | |
| CIENCIAS DE LA COMPUTACIÓN MATEMÁTICAS INDUSTRIALES | |
| The main goal of this work is exploring, evaluating, implementing and adapting some state-of-the-art models to evaluate executions of movements by karate athletes. In sports, the evaluation of individual motion skills plays a fundamental role. Unfortunately, human motion involves many correlated variables, both temporally and spatially. To this end, we study first the effects of different parameterizations, based on orientations and on positions and we evaluate each method's sensitivity to temporal distortions. Afterwards, we explore methods to evaluate human motion, starting with traditional approaches based on measurements using the raw data, like Linear Matching and non-Linear Matching approaches. Next, we discuss a method based on encoding the correct motion sequences through a probabilistic model. Finally, we explore some more complex methods based on deep learning. We implemented a supervised, unsupervised and deep metric learning model. | |
| 01-06-2021 | |
| Trabajo de grado, maestría | |
| OTRAS | |
| Versión aceptada | |
| acceptedVersion - Versión aceptada | |
| Aparece en las colecciones: | Tesis del CIMAT |
Cargar archivos:
| Fichero | Descripción | Tamaño | Formato | |
|---|---|---|---|---|
| TE 807.pdf | 68.26 MB | Adobe PDF | Visualizar/Abrir |