Modelo de interação natural com imagem projetada : proposta alternativa ao uso de sensores reais, baseado em algoritmos de IA e processamento de imagens

Abstract

The intuitive interaction between humans and computers is a field of research that has been investigated in recent years. The use of different types of sensors, such as motion sensors, touch, etc., have provided great strides in this area. However, according to the specificity of each sensor, the cost for developing some applications it is often infeasible. Currently with greater access to image capture devices and computers with enough processing capability, you can identify interaction without the need for special devices. Thus, this work has as main objective the proposal for a model as an alternative to using real sensors, in this case, touch sensors for human-machine interaction. As validation of the proposed model was developed an application that makes possible the interaction between the user and the image projected by an image projector. For this, we used the techniques of artificial intelligence and processing images, which allowed to identify relevant information from the scene (targets), isolating them through the use of image segmentation, and finally, the use of the term histogram to identify whether or not there was interaction.