Journal of Intelligent Computing Felipe José Aguiar Maia, José Everardo Bessa Maia, Thelmo Pontes de Araujo 9 4 2018 https://doi.org/10.6025/jic/2018/9/4/157-165 http://www.dline.info/jic/fulltext/v9n4/jicv9n4_3.pdf This work looks into the problem of detecting the presence of motorcyclist safety helmet on single images of city roads. The degree of adjustment of a new image to the previously trained Active Appearance Model (AAM) acts as the decision criterion for detection. This adjustment is measured by differences between AAM shape and appearance parameter vectors, which become the feature vectors for four different classifiers. Results show that this approach is feasible and may be recommended.