Towards a possibilistic classification of gastroenterology patterns in a complex environment

International audienceIn this paper, several types of imperfection and uncertainty that may affect the information elements in any pattern recognition system will be pondered, whether when object classification depends on datasets, or when it relies on expert knowledge sets. Afterwards, all the discussed imperfection aspects will be processed within a unified framework, fundamentally based on possibility theory and fuzzy relation composition rules. At last, various concrete illustrative examples applied to gastroenterology patterns will be given and discussed