Automatic classification of Nosema pathogenic agents through machine vision techniques and kernel-based vector machines

Over the past few years, the microscopic image analysis has become increasingly important for the diagnosis and classification of diseases in natural and health sciences. Although some computational tools are available for image processing on those areas, their efficiency is limited by lack of adaptation to the specific problem. This work presents a simple and direct method to identify and classify spores with the use of machine vision and supervised learning techniques in order to detect diseases in bee colonies. The method makes use of segmentation techniques to identify spores which are subsequently classified by means of multi-class kernel-based vector machines. Different computer vision tools have been combined and applied to enhance the images and get the relevant information. The results are encouraging and are also applicable to the diagnosis of other parasitic diseases.