Facial image acquisition systems produce low quality face images. This happens because the imaging conditions like illumination, occlusion or noise might change among images. To achieve optimal images, we proposed an image acquisition method for face recognition. Then, with this method, it was created the Smart Event Faces Database that contains video frames from videos taken by smartphones and Raspberry Pi. Also, it was measured the accuracy for face recognition and execution time for the Smart Event Faces Database using ResNet 34 for feature extraction and the next classifiers: K-Nearest Neighbors, Naive Bayes, Random Forest, Multi-Layer Perceptron, Decision Tree, Adaboost and Support Vector Machine. Additionally, we compared these classifiers to show which was effective for the dataset in terms of accuracy and execution time. Then, we used the Smart Event Faces Database to create an automatic attendance system for events using Raspberry Pi, ResNet-34 and K-Nearest Neighbors classifier. The results achieved in the Smart Event Faces Database showed that K-Nearest Neighbors and Support Vector Machine had the best results with more than 0.96 of accuracy for face recognition and less than 1.5 seconds respectively of execution time. The automatic attendance system had an accuracy for face recognition of 0.94 and 0.5 seconds approximately per frame in execution time for 19 persons in 2 events.