Addressing the problem of image classification with low-resolution images has been a focused research topic for many years. One of the major challenges of this issue arises from the limited information in the dataset. Most proposed solutions to this challenge involve altering the architecture of neural network models. In this paper, we introduce a novel approach that utilizes data augmentation through deep convolutional generative adversarial networks (DCGAN) to enrich the information learned by the model. Experimental results on the Animal Faces-HQ (AFHQ) dataset at the resolution 64 × 64 demonstrate that the DCGAN model effectively enriches the training dataset, leading to improvements in the classification model’s performance.