Avaliação de escoliose utilizando baropodômetro e rede neural artificial
Carregando...
Data
2017-01-16
Autores
Fanfoni, Caroline Meireles [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A patologia mais recorrente na coluna vertebral é a escoliose. A modificação estrutural causada pela escoliose gera o desalinhamento postural global do indivíduo. Uma das modificações causadas pelo desalinhamento postural é a forma como o indivíduo distribui o peso na região plantar dos pés. O objetivo deste trabalho foi implementar um sistema eletrônico constituído por um baropodômetro e redes neurais artificiais para separar pacientes com Grau I na classificação de Ricard, de 1o a 19o de escoliose, em dois grupos, C1 (1o a 9o) e C2 (10o a 19o). A maior percentagem de pacientes com escoliose está nesta faixa, aqueles que não precisam usar coletes ou fazer cirurgia, e cujo tratamento é realizado com ginásticas especiais e com avaliações freqüentes pelo profissional de saúde. A classificação dos pacientes nos grupos de escoliose foi implementada com o software Matlab e redes neurais artificiais, usando o algoritmo de treinamento backpropagation. A precisão média da classificação foi de 93,7% para o grupo C1 e 94,5% para o grupo C2. As acurácias na classificação foram de 83,3% para o grupo C1 e 96% para o grupo C2. O sistema implantado pode contribuir para o tratamento de pacientes com grau de escoliose na faixa de 1o a 19o, o intervalo de maior incidência desta patologia, no qual o monitoramento da condição clínica com técnicas não invasivas é de fundamental importância.
The most recurrent pathology in the spine is scoliosis. The structural modification caused by scoliosis generates the individual's overall postural misalignment. One of the modifications caused by postural misalignment is the way the individual distributes the weight in the plantar region of the feet. The objective of this work was to implement an electronic system consisting of a baropodometer and artificial neural networks to separate patients with Grade I in the classification of Ricard, from 1o to 19o of scoliosis, in two groups, C1 (1o to 9o) and C2 (10o to 9o). The largest percentage of patients with scoliosis are in this range, those who do not need to wear vests or have surgery, and whose treatment is performed with special gymnastics and with frequent evaluations by the health professional. The classification of patients in the scoliosis groups was implemented with Matlab software and artificial neural networks, using the backpropagation training algorithm. The mean precision of the classification was 93.7% for the C1 group and 94.5% for the C2 group. The accuracy in the classification was 83.3% for the group C1 and 96.0% for the group C2. The implemented system can contribute to the treatment of patients with grade of scoliosis in the range from 1o to 19o, the interval of higher incidence of this pathology, in which the monitoring of the clinical condition by non-invasive techniques is of fundamental importance.
The most recurrent pathology in the spine is scoliosis. The structural modification caused by scoliosis generates the individual's overall postural misalignment. One of the modifications caused by postural misalignment is the way the individual distributes the weight in the plantar region of the feet. The objective of this work was to implement an electronic system consisting of a baropodometer and artificial neural networks to separate patients with Grade I in the classification of Ricard, from 1o to 19o of scoliosis, in two groups, C1 (1o to 9o) and C2 (10o to 9o). The largest percentage of patients with scoliosis are in this range, those who do not need to wear vests or have surgery, and whose treatment is performed with special gymnastics and with frequent evaluations by the health professional. The classification of patients in the scoliosis groups was implemented with Matlab software and artificial neural networks, using the backpropagation training algorithm. The mean precision of the classification was 93.7% for the C1 group and 94.5% for the C2 group. The accuracy in the classification was 83.3% for the group C1 and 96.0% for the group C2. The implemented system can contribute to the treatment of patients with grade of scoliosis in the range from 1o to 19o, the interval of higher incidence of this pathology, in which the monitoring of the clinical condition by non-invasive techniques is of fundamental importance.
Descrição
Palavras-chave
Escoliose, Baropodômetro, Redes neurais artificiais, Retropropagação