• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  Bookmark and Share
Master's Dissertation
Full name
Camila Cocolo
Knowledge Area
Date of Defense
São Carlos, 2020
Perdoná, Gleici da Silva Castro (President)
Novelli, Cibele Maria Russo
Suzuki, Adriano Kamimura
Tinós, Renato
Title in Portuguese
Reconhecimento de movimentos de cães utilizando um acelerômetro e redes neurais artificiais
Keywords in Portuguese
Classificação de movimentos
Abstract in Portuguese
A classificação dos movimentos de cães utilizando dados de acelerômetro é uma área ainda pouco explorada no Brasil, mas de grande importância para o acompanhamento da saúde e bem estar destes animais. Este trabalho propõe um método de classificação de movimentação dos cães, a partir de um acelerômetro triaxial, e utilização de três arquiteturas de redes neurais artificiais: Rede Neural Convolucional (CNN), Rede Neural Convolucional associada a Long Short Term Memory (CNN-LSTM) e ConvLSTM. A metodologia foi desenvolvida instalando um pingente contendo o acelerômetro na coleira de 8 cachorros, que coletava dados em uma frequência de 10 Hz. Para avaliar o desempenho das redes neurais foi considerado o coeficiente de Matthews, que é um indicador muito utilizado na área de bioinformática. A arquitetura com melhor desempenho foi a ConvLSTM, que apresentou um coeficiente de Matthews de 0,79 no conjunto de teste.
Title in English
Recognition of dog movements using an accelerometer and artificial neural networks
Keywords in English
Movement classification
Abstract in English
Classification of dogs movements by using data collected from accelerometers is an area little explored in Brazil, but this is of great importance to monitor health and well-being of these animals. This work proposes a method to classify the movement of dogs using a triaxial accelerometer, and the use of three artificial neural network architectures: Convolutional Neural Network (CNN), Convolutional Neural Network associated with Long Short Term Memory ( CNN-LSTM) and ConvLSTM. The methodology was developed by installing a pendant that contains the accelerometer on the collar of 8 dogs, and it presents data collected at a frequency of 10 Hz. To evaluate the neural network performance the Matthews coefficient was considered, which is an widely used indicator in the area of bioinformatics. The best performing architecture was ConvLSTM, which had a Matthews coefficient of 0.79 on the test set
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
CamilaCocolo.pdf (2.83 Mbytes)
Publishing Date
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2023. All rights reserved.