No description available.

The thesis shows the development of DSP-based ¡§human head-tracker¡¨ system. It uses CCD camera to capture images, and detects in the image sequence. When someone interrupts, the system will lock on his head and shows the locked image on the LCD screen. The Head-tracker system includes three sub-systems : ¡§Motion Detector¡¨, ¡§Ellipse Algorism¡¨, and ¡§Visual Probability Data Association Filter¡¨. From the results of experiment, it can meet the expectation and gain good performance and robustness.

Orientador: José Celso Freire Junior / Resumo: Algumas necessidades específicas da Fisioterapia, área que estuda o Equilíbrio Postural e a Análise da Marcha Humana estão entre as principais razões para o desenvolvimento deste trabalho de pesquisa. Ele, tem por base uma plataforma de força, desenvolvida em 2016 por pesquisadores da UNESP e do Laboratório G-SCOP, da instituição Grenoble INP da França. A esta solução foram adicionadas duas interfaces eletrônicas. A primeira, com objetivo de fornecer informações sobre as forças de reação do solo, de um paciente que está sobre a plataforma e a segunda para controlar a estabilidade da plataforma, simulando dois níveis de dificuldades para o paciente se manter em equilíbrio. Aplicativos foram desenvolvidos para fazer a captura de dados e a análise gráfica dos resultados. Além de calcular informações sobre o deslocamento dos centros de pressão e de massa do paciente, eles asseguram o sincronismo entre os dados capturados por sensores de força, instalados na plataforma e um sensor de movimento Kinect. O tratamento dos dados do sensor de movimento possibilita a criação de um modelo tridimensional detalhado, na forma de um esqueleto digital, que apresenta a cinemática de várias articulações do corpo humano e o seu respectivo comportamento posicional, em função do tempo. O trabalho apresenta também um estudo comparativo entre o sistema desenvolvido e um ambiente de referência padronizado, que utiliza um sistema de filmagem, com várias câmeras infravermelho de alta definição, disponív... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Some specific needs of Physiotherapy, an area that studies the Postural Equilibrium and the Gait Analysis are among the main reasons for the development of this research work. It is based on a force platform, developed in 2016 by researchers from UNESP and the G-SCOP Laboratory of the Grenoble INP institution in France. Two electronic interfaces were added to this solution. The first one, to provide information about the ground reaction forces, from one patient on the platform and the second to control the stability of the platform, simulating two levels of difficulty for the patient to keep in balance. Applications were developed to capture data and graphically analyze the results. In addition to calculating information about the displacement of the patient's pressure and mass centers, they ensure synchronization between the data captured by force sensors installed on the platform and a Kinect motion sensor. The treatment of motion sensor data allows the creation of a detailed three-dimensional model, in the form of a digital skeleton, which presents the kinematics of various joints of the human body and their respective positional behavior, as a function of time. The work also presents a comparative study between the developed system and a standardized reference environment, using a multi-camera high-definition infrared camera system, available in a dedicated room of the University Hospital of Grenoble. This analysis evaluates the possibility of creating new products for th... (Complete abstract click electronic access below) / Doutor

近年來人體姿態感測的技術與應用愈來愈熱門。許多電玩及電影大多是藉由攝影鏡頭偵測人體姿態，但效果容易受到外在因素的影響，如陽光、遮蔽物等，特別是用在舞台展演方面就不適合了。因此本篇論文希望設計穿戴式裝置來感測人體姿態，並結合新一代無線網路藍牙4.0做為數據傳輸方式，以進行即時互動展演。本論文研究開發以Arduino結合六軸姿態感測器，設計此穿戴式裝置雛型。而實際在展場上表演，此裝置必需具有一定的穩定性、可靠性且方便穿戴。本論文設計出的裝置與系統，已實際參與數場即時互動展演，並且供觀眾親自體驗，反應熱烈。未來期許能將此系統擴大應用於多人異地互動展演上、能有更大的擴充性、多樣的互動體驗。 / Recently, human body skeleton detection technology and its applications are becoming more popular. Many computer games and movies detect human body skeleton by cameras. However, the detection will be affected easily by sun light or obstacles. Especially, this is not suitable for the applications on stage performance. The goal of this thesis is to design wearable devices to detect body skeleton and uses BLE 4.0 for data transmission to interactively real time art perform. We put 6-axis motion detector on Arduino to develop the wearable device prototyping. These devices for performing in practice should be stable, reliable and convenient for wearing or taking off. The devices and system we developed have been used in several real time interactive performances and for audience experience. In the future, this system can be expanded to people performing in different locations and has better scalability and varieties of interactive experience.