Detector de faces baseado no pós-processamento da detecção de pele.

Abstract

Detecting people in an image is one application of the Vision Computer and Digital Image Processing, and face detection is one of the most used algorithm to do this task. Face detection’s goal is to find in an image an area where the pixels corresponds to a human face. In order to achieve a better efficiency in face detection field, human skin features can be combined with a face detector to reduce the number of false detections of the classifier. In this work, it is presented an Face Detector Based on Skin Detection (FDBSD) algorithm that increases the precison (PPV) of the face detector and reduces the number of false detections. To do that, it was used a skin detector based on coloured histograms of Jones and Regh and the face detectors of Haar Cascade and PICO.