Research on batik image pattern detection based on improved YOLOv11

Batik, as an important intangible cultural heritage, embodies profound cultural significance through intricate pattern systems. However, detecting these patterns in complex batik images poses significant challenges due to dense pattern distributions, scale variations, complex backgrounds, and degraded image quality. This paper proposes a robust batik pattern detection model based on improved YOLOv11 architecture that balances detection accuracy with computational efficiency. First, we construct a comprehensive Chinese batik dataset, addressing the critical data scarcity in this domain. Second, we develop an enhanced YOLOv11 model integrating Vision Outlooker (VOLO) attention mechanisms for capturing long-distance spatial dependencies and Fused-MBConv modules for efficient feature extraction. Third, we implement a prototype system that bridges visual detection with cultural knowledge interpretation through batik knowledge graphs. The proposed approach provides a practical and scalable solution for the digital preservation and interpretation of intangible cultural heritage.