Identification of Geometric Shapes with Real-Time Neural Networks

This article presents the implementation of an identification system of geometric figures and their respective colors, this made with neural networks by using Backpropagation control implementation. This paper describes the process of extracting characteristic patterns of images, with the help of Artificial Neural Networks. The information Neuronal Network along with additional data of images and colors, will be stored in a database which will be put dynamic that will evolve with the figures they will be learning this by implementing a PID created in MATLAB Software. Subsequent to perform image capture with an independent PC WEBCAM. This processes the image and along with the data acquired and processed by the neural network in the pattern of shapes and colors. For image processing libraries MATLAB be used both in the implementation of a system acquisition by WEBCAM.