Invariant recognition of polychromatic images of Vibrio cholerae O1

Abstract

Cholera is an acute intestinal infectious disease. It has claimed many lives throughout history, and it continues to be a global health threat. Cholera is considered one of the most important emergence diseases due its relation with global climate changes. Automated methods such as optical systems represent a new trend to make more accurate measurements of the presence and quantity of this microorganism in its natural environment. Automatic systems eliminate observer bias and reduce the analysis time.

We evaluate the utility of coherent optical systems with invariant correlation for the recognition of Vibrio cholerae O1. Images of scenes are recorded with a CCD camera and decomposed in three RGB channels. A numeric simulation is developed to identify the bacteria in the different samples through an invariant correlation technique. There is no variation when we repeat the correlation and the variation between images correlation is minimum. The position-, scale-, and rotation-invariant recognition is made with a scale transform through the Mellin transform. The algorithm to recognize Vibrio cholerae O1 is the presence of correlation peaks in the green channel output and their absence in red and blue channels. The discrimination criterion is the presence of correlation peaks in red, green, and blue channels.