Nowadays, biometric identification is utilized in several applications especially in security system. One of the recently popular biometric identifications is person identification from ear because each person has a unique ear and it does not change overtime. In addition, we believe that not only side view ear image is useful in identifying a person, but a front view ear image is also useful. Hence, in this paper, we develop two convolutional neural networks (CNNs) schemes to recognize front view and side view human ear. From the blind test data set results, we found that the system based on front view images provides 84% correct. Meanwhile, the side view image-based system yields 80% correct classification on the same data set.