Attitudes, gestures, postures, facial expressions, and speech are considered to be channels for the transmission of human emotions. Visual interaction is a powerful communication tool for individuals. Even a simple spontaneous change in facial expression conveys feelings such as happiness, sadness, surprise, or anxiety. This present work proposes a solution that automatically recognizes the emotion shown on a given face. Thus, a solution based on an optimized deep convolutional neural network (OP-DCNN) is used to classify the following emotions: Happiness, Sadness, Anger, Disgust, Surprise, Neutrality, and Fear. The impact of optimization techniques like regularization and hyperparameter tuning are discussed in this work. It outperforms state-of-the-art results, with an accuracy of 98% on the CK+ dataset.