Coordinate Attention-Based Neck Feature Fusion for Enhanced YOLOv11 Oil Palm FFB Ripeness Detection

Reza Hanif Kurniawan; Sri Winarno

doi:10.29408/edumatic.v10i2.36243

Authors

Reza Hanif Kurniawan University Dian Nuswantoro https://orcid.org/0009-0001-7425-3380
Sri Winarno Universitas Dian Nuswantoro https://orcid.org/0000-0002-6801-2767

DOI:

https://doi.org/10.29408/edumatic.v10i2.36243

Keywords:

coordinate attention, neck feature fusion, oil palm ffb ripeness detection, yolov11

Abstract

Detecting the ripeness of oil palm fresh fruit bunches (FFB) directly on trees is complicated by factors such as partial visibility, occlusion, low contrast, image blur, and scale variation in plantation images. This study evaluated how the choice and positioning of attention mechanisms affect neck feature fusion in the lightweight YOLOv11n model. The baseline YOLOv11n was compared with SE-YOLOv11n, CBAM-YOLOv11n, and CA-YOLOv11n, all trained under the same conditions, with the attention mechanisms applied at the first and second up-sampling fusion features. Additionally, Coordinate Attention was tested in three different neck placement setups. CA-YOLOv11n recorded the highest recall, F1-score, mAP@50, and mAP@50–95, with values of 0.772, 0.773, 0.813, and 0.518, respectively, while SE-YOLOv11n achieved the highest precision at 0.794. Compared with the baseline, CA-YOLOv11n added 4,640 parameters, increased GFLOPs from 6.4 to 6.5, and decreased inference throughput from 15.1 to 14.7 FPS. The placement ablation revealed that integrating attention at the first and second fusion features yielded the highest F1-score and mAP values among the three placements tested, whereas extending CA to all three fusion features diminished the detection performance. These findings suggest that within the YOLOv11n configuration assessed, both the selection and placement of attention mechanisms impact the detection performance, and adding more attention insertions does not necessarily enhance the results.

References

Ali, M. L., & Zhang, Z. (2024). The YOLO Framework: A Comprehensive Review of Evolution, Applications, and Benchmarks in Object Detection. Computers, 13(12), 336. https://doi.org/10.3390/computers13120336

An, X., Sun, L., Zhao, Y., Wang, H., Zhang, M., Li, K., & Xie, R. (2026). CHCG-YOLO: a lightweight model for fruit recognition. Journal of Food Measurement and Characterization, 20(2), 1971–1985. https://doi.org/10.1007/s11694-025-03775-9

Bian, Z., Yao, B., & Li, Q. (2025). SPAFPN: Wear a multi-scale feature fusion scarf around neck for real-time object detector. Information Fusion, 119, 103034. https://doi.org/10.1016/j.inffus.2025.103034

Chen, S., Liu, Y., Zhang, H., & Cai, Y. (2025). A human location and action recognition method based on improved Yolov11 model. Discover Artificial Intelligence, 5(1), 232. https://doi.org/10.1007/s44163-025-00492-6

Gong, X., & Wu, Q. (2025). Fruit Detection Methods Based on Deep Learning in Agricultural Planting: A Systematic Literature Review. IEEE Access, 13, 96092–96110. https://doi.org/10.1109/ACCESS.2025.3573364

Gundersen, O. E., Shamsaliei, S., Kjærnli, H. S., & Langseth, H. (2023). On Reporting Robust and Trustworthy Conclusions from Model Comparison Studies Involving Neural Networks and Randomness. Proceedings of the 2023 ACM Conference on Reproducibility and Replicability, 37–61. https://doi.org/10.1145/3589806.3600044

Hanafy, A. El, Hessane, A., & Farhaoui, Y. (2026). Exploring attention mechanisms in deep learning for plant health: A scoping review on disease and pest detection. Information Processing in Agriculture. https://doi.org/10.1016/j.inpa.2026.05.006

Hou, Q., Zhou, D., & Feng, J. (2021). Coordinate Attention for Efficient Mobile Network Design. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 13708–13717. https://doi.org/10.1109/CVPR46437.2021.01350

Junior, F. A., & Suharjito. (2023). Video based oil palm ripeness detection model using deep learning. Heliyon, 9(1), e13036. https://doi.org/10.1016/j.heliyon.2023.e13036

Lai, J. W., Ramli, H. R., Ismail, L. I., & Hasan, W. Z. W. (2022). Real-Time Detection of Ripe Oil Palm Fresh Fruit Bunch Based on YOLOv4. IEEE Access, 10, 95763–95770. https://doi.org/10.1109/ACCESS.2022.3204762

Li, H., Hu, Z., Wu, F., Li, J., Meng, H., Yang, H., & Wang, Z. (2026). A Study on a Method for Detecting Leaf Diseases in Pumpkins Based on an Improved YOLO11 Model. AgriEngineering, 8(7), 287. https://doi.org/10.3390/agriengineering8070287

Li, J., Zhang, T., Zeng, S., Gao, Q., Wang, L., & Lu, J. (2025). StarNet-Embedded Efficient Network for On-Tree Palm Fruit Ripeness Identification in Complex Environments. Agriculture, 15(17), 1823. https://doi.org/10.3390/agriculture15171823

Liu, S., Shao, F., Chu, W., Dai, J., & Zhang, H. (2025). An Improved YOLOv8-Based Lightweight Attention Mechanism for Cross-Scale Feature Fusion. Remote Sensing, 17(6), 1044. https://doi.org/10.3390/rs17061044

Liu, X., Li, G., Chen, W., Liu, B., Chen, M., & Lu, S. (2022). Detection of Dense Citrus Fruits by Combining Coordinated Attention and Cross-Scale Connection with Weighted Feature Fusion. Applied Sciences, 12(13), 6600. https://doi.org/10.3390/app12136600

Naftali, M. G., Hugo, G., & Suharjito. (2024). Palm Oil Counter: State-of-the-Art Deep Learning Models for Detection and Counting in Plantations. IEEE Access, 12, 90395–90417. https://doi.org/10.1109/ACCESS.2024.3419835

Sapkota, R., Meng, Z., Churuvija, M., Du, X., Ma, Z., & Karkee, M. (2026). Comprehensive performance evaluation of YOLOv12, YOLO11, YOLOv10, YOLOv9 and YOLOv8 on detecting and counting fruitlet in complex orchard environments. Agriculture Communications, 4(1), 100125. https://doi.org/10.1016/j.agrcom.2026.100125

Suharjito, Naftali, M. G., & Hugo, G. (2024). Oil Palm Fruit Dataset on Plantations for Harvesting Estimation (Version 3) [Dataset]. Science Data Bank. https://doi.org/10.57760/sciencedb.12647

Suharjito, Naftali, M. G., Hugo, G., Priyadi, M. R. A., Asrol, M., & Utama, D. N. (2025). Oil Palm Fruits Dataset in Plantations for Harvest Estimation Using Digital Census and Smartphone. Scientific Data, 12(1), 972. https://doi.org/10.1038/s41597-025-05227-x

Suparto, A., & Pribadi, M. R. (2026). Improving Oil Palm Fruit Detection under Class Imbalance Using Class-Balanced Focal Loss on YOLOv11. Jurnal Sisfokom (Sistem Informasi Dan Komputer), 15(02), 165–176. https://doi.org/10.32736/sisfokom.v15i02.2568

Ullah, M. R., Chuang, H., Kayes, A., & Rouf, M. A. (2025). Real-Time Aerial Object Detection via YOLOv11 with Coordinate Attention Enhancement. Proceedings of the 2025 3rd International Conference on Machine Learning and Pattern Recognition, 22–27. https://doi.org/10.1145/3760678.3760682

Wu, J., Dai, G., Zhou, W., Zhu, X., & Wang, Z. (2024). Multi-scale feature fusion with attention mechanism for crowded road object detection. Journal of Real-Time Image Processing, 21(2), 29. https://doi.org/10.1007/s11554-023-01409-1

Xie, F., Lin, B., & Liu, Y. (2022). Research on the Coordinate Attention Mechanism Fuse in a YOLOv5 Deep Learning Detector for the SAR Ship Detection Task. Sensors, 22(9), 3370. https://doi.org/10.3390/s22093370

Xu, J., Zhu, X., Shi, L., Li, J., & Guo, Z. (2024). Squeeze-and-excitation attention and bi-directional feature pyramid network for filter screens surface detection. Journal of Electronic Imaging, 33(04). https://doi.org/10.1117/1.JEI.33.4.043044

Xu, M., Zhang, X., Ma, N., & Li, Y. (2024). Detection method of corn leaf diseases based on CA-YOLOv8. INMATEH Agricultural Engineering, 129–138. https://doi.org/10.35633/inmateh-73-11

Zhao, Y., Liu, Q., Su, H., Zhang, J., Ma, H., Zou, W., & Liu, S. (2024). Attention-Based Multiscale Feature Fusion for Efficient Surface Defect Detection. IEEE Transactions on Instrumentation and Measurement, 73, 1–10. https://doi.org/10.1109/TIM.2024.3372229

Zhou, Y., Huang, W., Liu, B., Chen, T., Wang, J., Zhang, Q., & Yang, T. (2026). Lightweight GAFNet Model for Robust Rice Pest Detection in Complex Agricultural Environments. AgriEngineering, 8(1), 26. https://doi.org/10.3390/agriengineering8010026