Modification of YOLOv11 Architecture with Convolutional Block Attention Module Integration for Batik Besurek Motif Detection

Reksi Hendra Pratama; Ernawati; Widhia Oktoeberza KZ

doi:10.34148/teknika.v15i1.1449

Authors

Reksi Hendra Pratama Informatics Engineering, Faculty of Engineering, Universitas Bengkulu, Bengkulu, Indonesia
Ernawati Informatics Engineering, Faculty of Engineering, Universitas Bengkulu, Bengkulu, Indonesia
Widhia Oktoeberza KZ Informatics Engineering, Faculty of Engineering, Universitas Bengkulu, Bengkulu, Indonesia

DOI:

https://doi.org/10.34148/teknika.v15i1.1449

Keywords:

Batik Besurek, Object Detection, YOLOv11, CBAM, Deep Learning

Abstract

Batik Besurek is a distinctive cultural heritage textile from Bengkulu, characterized by the combination of Arabic calligraphy, floral, and faunal motifs. However, public recognition of these motifs continues to decline, particularly among younger generations, creating an urgent need for digital preservation through modern computer vision techniques. This study aims to develop an enhanced object detection model for identifying Batik Besurek motifs by integrating the Convolutional Block Attention Module (CBAM) into the Neck (feature fusion layers) of the YOLOv11 architecture. A total of 3,876 images were collected from online repositories and field documentation in Bengkulu, followed by annotation and data augmentation using the Roboflow platform. Specifically, Mosaic augmentation combined with a close-mosaic strategy proved most significant in stabilizing the detection of dense and minority motifs. The research employs an experimental methodology by constructing two models, standard YOLOv11 as the baseline and YOLOv11+CBAM as the proposed model, trained under identical hyperparameter configurations. Model performance was evaluated using standard object detection metrics, namely precision, recall, mAP@50, and mAP@50–95. The experimental results demonstrate that the integration of CBAM significantly improves feature extraction capability, enabling the model to better capture dense, small, and visually complex batik motifs. The proposed model outperforms the baseline across all metrics, achieving a mAP@50 of 0.988, mAP@50-95 of 0.914, Recall of 0.971, and Precision of 0.959. These results indicate superior sensitivity and localization accuracy, particularly for minority motif classes, confirming that incorporating attention mechanisms into YOLOv11 provides a robust solution for the digital preservation of Besurek patterns.

Downloads

Download data is not yet available.

References

[1] M. Igiriza, R. A. Valentino, G. Wiradharma, Z. Adamy, and A. Anjani, “Preservation of Balinese batik knowledge as an effort to preserve indigenous knowledge,” J. Kaji. Inf. Perpust., vol. 12, no. 2, pp. 205–220, 2024, [Online]. Available: https://jurnal.unpad.ac.id/jkip/article/view/58663

[2] E. N. Intania, “Revitalisasi Batik sebagai Simbol Identitas Nasional dalam Era Globalisasi dan Industri Kreatif Global,” Maliki Interdiscip. J., vol. 3, pp. 1050–1055, 2025, [Online]. Available: https://urj.uin-malang.ac.id/index.php/mij/article/view/14909

[3] M. Utami, “Model ResNet50 dengan Optimalisasi Activation Function dan Optimizer Berbasis Gamma -MSR Untuk Klasifikasi Citra Motif Batik,” Universitas Sriwijaya, 2025.

[4] Y. Rao et al., “Feature enhancement and attention mechanism fusion method for Cantonese embroidery image detection based on YOLOv8,” npj Herit. Sci., vol. 13, no. 1, p. 521, 2025, doi: 10.1038/s40494-025-02079-x.

[5] W. Azizah and S. Agustin, “Feature Extraction using Histogram of Oriented Gradients and Moments with Random Forest Classification for Batik Pattern Detection,” J. SISFOKOM (Sistem Inf. dan Komput., vol. 14, pp. 8–14, 2025, doi: 10.32736/sisfokom.v14i1.2352.

[6] E. Khoirunnisa, F. Alzami, R. A. Pramunendar, and R. A. Megantara, “Enhanced Semarang Batik Classification using MobileNetV2 and Data Augmentation,” Sinkron, vol. 9, no. 1, pp. 43–54, 2025, doi: 10.33395/sinkron.v9i1.14308.

[7] A. F. Rozi, A. Wibowo, and B. Warsito, “Resampling Technique for Imbalanced Class Handling on Educational Dataset,” JUITA, vol. 11, no. 1, pp. 77–85, 2023, doi: 10.30595/juita.v11i1.15498.

[8] S. A. Sutresno, “The Classification of Batik Besurek Fabric Motifs in Indonesia Utilizing YOLOv8 for Enhanced Cultural Preservation,” J. Comput. Syst. Informatics, vol. 6, no. 1, pp. 86–95, 2024, doi: 10.47065/josyc.v6i1.6123.

[9] M. Fajar, J. Permana, J. Caesar, R. Bakar, and N. A. Fauzan, “YOLOv11 Model as a Smart Solution for Waste Identification and Classification in Automated Waste Management System,” J. Ilmu Komput. dan Inf., vol. 2, pp. 207–216, 2025, doi: 10.21609/jiki.v18i2.1490.

[10] Y. Wang et al., “A lightweight YOLOv11-based framework for small steel defect detection with a newly enhanced feature fusion module,” Sci. Rep., vol. 15, no. 1, p. 34322, 2025, doi: 10.1038/s41598-025-16619-9.

[11] M. Mao and M. Hong, “YOLO Object Detection for Real-Time Fabric Defect Inspection in the Textile Industry: A Review of YOLOv1 to YOLOv11,” Sensors, vol. 25, no. 7, pp. 1–43, 2025, doi: 10.3390/s25072270.

[12] Q. Li, T. Wu, T. Xu, J. Lei, and J. Liu, “A Novel YOLO Algorithm Integrating Attention Mechanisms and Fuzzy Information for Pavement Crack Detection,” Int. J. Comput. Intell. Syst., vol. 18, no. 1, p. 158, 2025, doi: 10.1007/s44196-025-00894-5.

[13] L. Z. Id, B. Ma, Y. Dong, Z. Yin, and F. Lu, “DCFE-YOLO : A novel fabric defect detection method,” pp. 1–21, 2025, doi: 10.1371/journal.pone.0314525.

[14] S. Woo, J. Park, J.-Y. Lee, and I. S. Kweon, “CBAM: Convolutional Block Attention Module,” in Computer Vision -- ECCV 2018, V. Ferrari, M. Hebert, C. Sminchisescu, and Y. Weiss, Eds., Cham: Springer International Publishing, 2018, pp. 3–19. doi: 10.1007/978-3-030-01234-2_1.

[15] H. Chen, Z. Chen, and H. Yu, “Enhanced YOLOv5 : An Efficient Road Object Detection Method,” Sensors, vol. 23, no. 20, 2023, doi: 10.3390/s23208355.

[16] S. Praveen and Y. Jung, “CBAM-STN-TPS-YOLO : Enhancing Agricultural Object Detection through Spatially Adaptive Attention Mechanisms,” 2023, doi: 10.48550/arXiv.2506.07357.

[17] N. Garshasebi, “Improving Small Object Detection in Remote Sensing with YOLO11- CBAM and Deep Learning,” vol. 50, no. 1, pp. 217–227, 2025, doi: 10.5281/zenodo.15845503.

[18] R. Sapkota and M. Karkee, “Integrating YOLO11 and Convolution Block Attention Module for Multi-Season Segmentation of Tree Trunks and Branches in Commercial,” pp. 1–19, 2024, doi: 10.48550/arXiv.2412.05728.

[19] N. Putu, D. Ariani, S. Dewi, J. Aji, D. Aryasa, and I. G. Hendrayana, “YOLOv8-Based Quality Detection of Bali MSMEs Staple Food,” TIERS, vol. 6, no. 2, pp. 171–186, 2026, doi: 10.38043/tiers.v6i2.7144.

[20] Y. Jin et al., “A Real-Time Fabric Defect Detection Method Based on Improved YOLOv8,” Appl. Sci., vol. 15, no. 6, 2025, doi: 10.3390/app15063228.

[21] M. B. BESUREK, “Object Detection Motif Asli Batik Besurek Dataset,” Mar. 2024, Roboflow. [Online]. Available: https://universe.roboflow.com/motif-batik-besurek/object-detection-motif-asli-batik-besurek

[22] S. Nepal et al., “MoodCapture: Depression Detection Using In-the-Wild Smartphone Images.,” Proc. SIGCHI Conf. Hum. factors Comput. Syst. CHI Conf., vol. 2024, May 2024, doi: 10.1145/3613904.3642680.

[23] T. Kumar, R. O. B. Brennan, S. Member, A. Mileo, and M. Bendechache, “Image Data Augmentation Approaches : A Comprehensive Survey and Future Directions,” 2023, doi: 10.48550/arXiv.2301.02830.

[24] T. Islam, M. S. Hafiz, J. R. Jim, M. M. Kabir, and M. F. Mridha, “A systematic review of deep learning data augmentation in medical imaging: Recent advances and future research directions,” Healthc. Anal., vol. 5, p. 100340, 2024, doi: https://doi.org/10.1016/j.health.2024.100340.

[25] M. H. Putri, U. Zaky, and B. A. Prabawa, “Optimizing Data Augmentation Parameters in YOLOv11 for Enhanced Rip Current Detection on Small Datasets from Depok-Parangtritis Coastline,” J. Tek. Inform., vol. 6, no. 5, pp. 3938–3957, 2025, doi: 10.52436/1.jutif.2025.6.5.5352.

[26] S. Chen, Y. Liu, H. Zhang, and Y. Cai, “A human location and action recognition method based on improved Yolov11 model,” Discov. Artif. Intell., vol. 5, no. 1, p. 232, 2025, doi: 10.1007/s44163-025-00492-6.

[27] R. R. Hamidi, M. K. Herliansyah, D. Sukma, and E. Atmaja, “Comparative Analysis of YOLOv5n and YOLOv8n Deep Learning Models for Precision Detection of Klowong Defects in Batik Fabric,” TIERS, vol. 6, no. 1, pp. 74–86, 2025, doi: 10.38043/tiers.v6i1.6499.

[28] J. Gao and Z. Wang, “YOLOv8-CBAM : a study of sheep head identification in Ujumqin sheep,” no. February, pp. 1–10, 2025, doi: 10.3389/fvets.2025.1514212.

[29] S. N. Rao, “YOLOv11 Architecture Explained: Next-Level Object Detection with Enhanced Speed and Accuracy,” www.medium.com. Accessed: Jan. 20, 2026. [Online]. Available: https://medium.com/@nikhil-rao-20/yolov11-explained-next-level-object-detection-with-enhanced-speed-and-accuracy-2dbe2d376f71

[30] W. Cai, Y. Chen, X. Qiu, M. Niu, and J. Li, “LLD-YOLO : A Low-Light Object Detection Algorithm Based on Dynamic Weighted Fusion of Shallow and Deep Features,” IEEE Access, vol. 13, no. January, pp. 69967–69979, 2025, doi: 10.1109/ACCESS.2025.3558574.

[31] I. Goodfellow, Deep Learning. 2016.

[32] X. Wang et al., “Research on multi class pests identification and detection based on fusion attention mechanism with Mask-RCNN-CBAM,” Front. Agron., vol. Volume 7-2025, 2025, doi: 10.3389/fagro.2025.1578412.