Aplicación para el procesamiento digital de imágenes de rostros

Abstract

The present work is about the development of an application for the digital processing of images of faces. On the other hand, due to the complexity of the characteristics of a digital image of a face, an architecture is proposed for the automation of the identification process. The process is divided in 3 stages, in the first one a video of 15 seconds of duration is obtained, where the application obtains a sequence of photos and processes them filtering the images that contain a clear face and that can clearly identify the facial descriptors such as eyes, nose and mouth. All these images are saved with a folder structure name of the person and all photos are evaluated against the 68 reference points of a face. In the second stage, the main features of the image are identified using cascade classification algorithms and the knn algorithm. Finally, in the third stage it is possible to detect and recognize the face. The results show an accuracy of 86.6%.