Evaluating the Similarity of Mobile App UI Designs: An Integrated Approach Using User Perception and AI
Hayun Park, Hyun K. Kim · International Journal of Human-Computer Interaction · 2026
As mobile app competition intensifies, differentiating user experience through creative UI design is vital. However, quantitative methodologies for evaluating UI similarity remain scarce. In this study, the key elements of UI similarity (layout, color, and content) are derived through user surveys and a similarity calculation algorithm is proposed using deep learning techniques based on YOLOv8 and convolutional autoencoder. We calculate layout similarity using intersection over union and feature vectors, color similarity using K-means and CIEDE2000, and content similarity using optical character recognition and SRoBERTa. The results of the comparison between the developed algorithm output and the user-rated similarity values showed a high correlation between the two, demonstrating the validity of the algorithm. This method not only contributes to ensuring objectivity and consistency in UI similarity evaluation but is also expected to be practically applicable in preventing design plagiarism and promoting the development of innovative UI designs.