Adaptive Methodology to Daytime and Night Environments for Eye and Mouth Detection Based on Artificial Vision

Abstract

Computer vision plays an important role in problem solving, mainly in those problems which are focused on human monitoring, being the facial features the most important among them. In this paper, a methodology for detecting eyes and mouth via software is presented, using image processing techniques such as color space conversion, thresholding and erosion. It has the characteristic of being adaptable to any person and works in both day and night environments, taking advantage in the ability to detect which environment is. To obtain the image in a daytime environment, the image is captured with natural light, while for night using an infrared capture system becomes. The system was tested in a software application developed with satisfactory results.