<p>This paper presents a multimedia-enabled deep learning framework for classifying and interpreting the brand styles of computer mouse designs. Leveraging a Convolutional Neural Network (CNN) based on VGG16, the system is trained on a labeled dataset of 6837 images spanning two contrasting brands. We integrate Class Activation Maps (CAMs) to visualize and interpret the style-relevant visual cues that influence brand predictions. To further understand model output, t-SNE and PCA analyses are employed to explore stylistic differentiation. In addition to technical evaluation, the system is validated through perceptual feedback collected from design students, bridging computational results with human-centered interpretation. This research contributes to the domain of multimedia tools for visual brand identity analysis, with potential applications in advertising, product design education, and digital design consulting platforms. This study proposes a data-driven framework for analyzing and predicting brand styles based on product form. Using convolutional neural networks (CNN), dimensional reduction techniques (t-SNE and PCA) and heatmap visualization, we examine visual identity patterns in two computer mouse brands—Genius and Mad Catz. The objective is to provide brand and design managers with scalable and interpretable tools for evaluating the coherence and differentiation of brand styles through automated visual analysis. Our method not only classifies brand styles with moderate precision, but also offers explicable outputs that highlight key shape features contributing to classification. A perceptual validation study involving 118 design students further assesses the alignment between artificial intelligence predicted classifications and human interpretations of the brand style. Results indicate that a explication AI can support both brand identity assessment and design education by providing insight into stylistic coherence. This research contributes to the growing literature on artificial intelligence in branding by demonstrating how machine learning can help maintain brand consistency, support design quality control and facilitate brand strategy communication. The findings have implications for product portfolio management, designer training and brand identity audit across industries.</p>