<p>The purpose of this study is to develop a novel convolutional neural network (CNN) concatenated with color histograms capable of identifying igneous rocks. General rock classification methodologies using deep learning relied only on CNN to analyze pattern features of images. However, recognizing the significance of mineral distribution in rock classification, additional layers for color features corresponding to different minerals were concatenated into the deep learning model. To validate this proposed model, 12 types of igneous rocks with diverse textures and colors were selected. Data augmentation techniques were applied to expand this dataset to 12,000 microscope images for training. For the reference model, CNN-only models, such as DenseNet, Inception, and Xception were used. Additionally, three experiments were conducted with the proposed method, integrating normal, inverted, and stretched color histograms into the CNN model. The results demonstrated that the concatenation of color histograms and CNN significantly enhanced classification performance. Specifically, the Xception model combined with a normal color histogram achieved the highest accuracy of 96.3%. For InceptionV3, the integration of an inverted color histogram yielded the most notable improvement, increasing accuracy by 6.3 percentage points. Overall, the addition of color histograms improved various performance indicators, including accuracy, recall, precision, F1-score, and kappa coefficient, compared to models that utilized only CNNs.</p>