Human Activity Recognition (HAR) is an interdisciplinary research area that has been attracting interest from several research communities specialized in machine learning, computer vision, medical and gaming research. The potential applications range from surveillance systems, human computer interfaces, sports video analysis, digital shopping assistants, video retrieval, games and health-care. In order to recognize a human action, the typical approach is to use manually labeled data to perform supervised training. This paper aims to compare the performance of several supervised classifiers trained with manually labeled data versus the same classifiers trained with data automatically labeled. The application should recognize an action performed in a sequence of continuous actions recorded with a Kinect sensor that provides the position of the main skeleton joints. In this paper we propose a framework capable of recognizing human actions using supervised classifiers trained with automatically labeled data.

--

machine learning,supervised learning,classification

• Ciências da Computação e da Informação - Ciências Naturais