Recognizing emotions in facial expressions computer is an interest- ing and challenging problem that can be applied in different contexts as tools to provide virtual personalization settings based on the character- istics of affective system user. This paper presents a study of automatic detection techniques of facial emotions that can be easily incorporated in virtual environments, ranking seven different emotional states: neutrality, happiness, sadness, surprise, anger, disgust and fear, through geometric approach and analysis representation of the image, making tagging and rating points of interest (landmarks) of the face. This work will also be presented the advances, difficulties encountered and future proposals to improve the Diagnostic accuracy of results