Voting and Ensemble Schemes Based on CNN Models for Photo-Based Gender Prediction

Gender prediction accuracy increases as convolutional neural network (CNN) architecture evolves. This paper compares voting and ensemble schemes to utilize the already trained five CNN models to further improve gender prediction accuracy. The majority voting usually requires oddnumbered models while the proposedsoftmaxbased voting can utilize any number of models to improve accuracy. The ensemble of CNN models combined with one more fullyconnected layer requires further tuning or training of the models combined. With experiments, it is observed that the voting or ensemble of CNN models leads to further improvement of genderprediction accuracy and that especially softmaxbased voters always show better gender prediction accuracy than majority voters. Also, compared with softmaxbased voters, ensemble models show a slightly better or similar accuracy with added training of the combined CNN models. Softmaxbased voting can be a fast andefficient way to get better accuracy without further training since the selection of the top accuracy models among available CNN pretrained models usually leads to similar accuracy to that of the corresponding ensemble models. KCI Citation Count: 0