Desenvolvimento de um programa computacional para análise técnico-tática no kendô

• Ano de defesa: 2019
• Autor(a) principal: Louzada, Ian Renon
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Educação Física; Centro de Educação Física e Desportos; UFES; Programa de Pós-Graduação em Educação Física
• Palavras-chave em Português: Scout; kendô; esportes de combate; análise técnica; Análise tática; Educação Física; 796
• Link de acesso: http://repositorio.ufes.br/handle/10/11277

The martial arts have had great repercussion in recent years due to their sportivization, potentialized by the entry of several modalities in the Olympic program, such as judo, taekwondo and karate. Kendo is a very traditional martial art, where practitioners and organizers try to keep it off the sports circuit, not giving it the sporting character. Despite this, there is development in competitions, finding a scientific way to try to improve fighting techniques. The present study aims to develop a methodology of technical and tactical analysis in Kendo. To this objective, a video analysis software was developed in order to facilitate the collection of data from the fights, having been named VideoK. The fights analyzed were the 15th Japanese Kendo Championship and were made available on a video streaming site by the Japanese Kendo Federation itself. An evaluator conducted two collections of the fifteen selected championship fights to evaluate the intra-evaluator correlation. Correlations classified as "very good", indicating good reproducibility, were obtained in all analyzes. It was also possible to obtain the averages of the targets of the fifteen fights, with the highest occurrence being the Men (head), followed by Kote (hand), Do (trunk) and Tsuki (throat). The interevaluator reproducibility was also analyzed, where a second evaluator collected data from ten fights of the same championship and the data were compared with the first evaluator's collection. All correlations were classified as "very good" between the two evaluators.