Huarui Liu
ABSTRACT
Although the target detection technology has made amazing progress in the detection of large and medium objects, due to the limited size of small objects, the occlusion between targets and the limitations of the convolutional network itself, small target detection is still as a challenging. In this paper, by improving the YOLOv5s algorithm, a lightweight small target detection FL-YOLO model with a hybrid structure is proposed, and the standard convolution in the original backbone network is replaced by depth-separable convolution and the residual connection method is adjusted. lightweight. A Patch-Attention (PA) module is proposed to extract context information of small objects and insert it into the backbone to enhance the feature extraction ability for tiny target. A new upsampling module named “Transition” is proposed to replace the nearest interpolation in YOLOv5s to minimize information conflict and information redundancy. Finally, the public dataset VisDrone2019 specially constructed for small targets was used to conduct experiments, which verified that the FL-YOLO proposed in this paper is more effective than YOLOv5s.
CCS Concepts • Computing methodologies∼Visual inspection
- Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).Google Scholar
Digital Library
- Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 779-788).Google ScholarCross Ref
- Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C. Y., & Berg, A. C. (2016). Ssd: Single shot multibox detector. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14 (pp. 21-37). Springer International Publishing.Google Scholar
- Noh, J., Bae, W., Lee, W., Seo, J., & Kim, G. (2019). Better to follow, follow to be better: Towards precise supervision of feature super-resolution for small object detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 9725-9734)Google ScholarCross Ref
- Yang, C., Huang, Z., & Wang, N. (2022). Querydet: Cascaded sparse query for accelerating high-resolution small object detection. In Proceedings of the IEEE/CVF Conference on computer vision and pattern recognition (pp. 13668-13677).Google ScholarCross Ref
- Liu, S., Qi, L., Qin, H., Shi, J., & Jia, J. (2018). Path aggregation network for instance segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 8759-8768).Google ScholarCross Ref
- Ghiasi, G., Lin, T. Y., & Le, Q. V. (2019). Nas-fpn: Learning scalable feature pyramid architecture for object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 7036-7045).Google ScholarCross Ref
- Liang, Z., Shao, J., Zhang, D., & Gao, L. (2018). Small object detection using deep feature pyramid networks. In Advances in Multimedia Information Processing–PCM 2018: 19th Pacific-Rim Conference on Multimedia, Hefei, China, September 21-22, 2018, Proceedings, Part III 19 (pp. 554-564). Springer International Publishing.Google ScholarDigital Library
- Tan, M., Pang, R., & Le, Q. V. (2020). Efficientdet: Scalable and efficient object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 10781-10790).Google ScholarCross Ref
- Xu, C., Wang, J., Yang, W., Yu, H., Yu, L., & Xia, G. S. (2022, November). RFLA: Gaussian receptive field based label assignment for tiny object detection. In Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part IX (pp. 526-543). Cham: Springer Nature Switzerland.Google Scholar
- Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., ... & Guo, B. (2021). Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 10012-10022).Google ScholarCross Ref
- Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., ... & Adam, H. (2017). Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861.Google Scholar
- Vaswani A, Shazeer N, Parmar N, Attention is all you need[J]. Advances in neural information processing systems, 2017, 30.Google Scholar
- Dosovitskiy A, Beyer L, Kolesnikov A, An image is worth 16x16 words: Transformers for image recognition at scale[J]. arXiv preprint arXiv:2010.11929, 2020.Google Scholar
- Zhang H, Li F, Liu S, Dino: Detr with improved denoising anchor boxes for end-to-end object detection[J]. arXiv preprint arXiv:2203.03605, 2022.Google Scholar
- Lin, T. Y., Goyal, P., Girshick, R., He, K., & Dollár, P. (2017). Focal loss for dense object detection. In Proceedings of the IEEE international conference on computer vision (pp. 2980-2988).Google ScholarCross Ref
- Law, H., & Deng, J. (2018). Cornernet: Detecting objects as paired keypoints. In Proceedings of the European conference on computer vision (ECCV) (pp. 734-750).Google ScholarDigital Library
- Bochkovskiy A, Wang C Y, Liao H Y M. Yolov4: Optimal speed and accuracy of object detection[J]. arXiv preprint arXiv:2004.10934, 2020.Google Scholar
- Duan, K., Bai, S., Xie, L., Qi, H., Huang, Q., & Tian, Q. (2019). Centernet: Keypoint triplets for object detection. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 6569-6578).Google ScholarCross Ref
Index Terms
- FL-YOLO: Lightweight Small Target Detection Algorithm Based on Transformer and CNN Hybrid Structure
Recommendations
UAV small target detection algorithm based on an improved YOLOv5s model
AbstractThe targets of UAV target detection are usually small targets, and the backgrounds are complex. In this work, aiming at the problem that small targets are easy to be missed or misdetected during the UAV detection, an improved YOLOv5s_MSES target ...
Highlights- Improved YOLOv5s_MSES algorithm for small target detection in UAV aerial photography.
- Introduction of Small target detection module STD to enhance detection of small targets.
- Multi-scale feature fusion module for improved accuracy ...
Infrared dim and small target detection based on U-Transformer
AbstractInfrared dim and small target detection is a key technology for space-based infrared search and tracking systems. Traditional detection methods have a high false alarm rate and fail to handle complex background and high-noise ...
Highlights- Based on the U-shaped network, the feature extraction ability of the detector to the target is improved.
Oriented Target Detection Algorithm Based on Transformer
AIPR '21: Proceedings of the 2021 4th International Conference on Artificial Intelligence and Pattern RecognitionOriented target detection is an important task in the field of remote sensing target detection, which has great application prospects in geography, agriculture and military. However, the current popular algorithms of oriented target detection do not ...
Comments