针对现有的舰船目标分类方法对舰船细粒度分类性能不佳、舰船图像特征学习效果差的问题，提出一种深度特征协作的舰船目标分类算法。首先，搭建双分支ResNet-18网络结构；然后引入对比学习的思想，实现双分支特征信息互补，丰富舰船图像特征学习；最后，通过特征协作模块，对学习到的双分支对比特征进行深度信息整合，以最小化分类损失，进而提高分类结果。在舰船图像数据集FGSC-23上的大量实验结果表明，对23类细粒度舰船图像分类平均准确率达到83.56%。
A deep feature collaborative ship target classification algorithm is proposed to address the issues of poor performance in fine-grained ship classification and poor learning of ship image features in existing ship target classification methods. Firstly, build a dual branch ResNet-18 network structure; Then, the idea of contrastive learning is introduced to achieve complementary feature information between two branches, enriching the learning of ship image features; Finally, through the feature collaboration module, the learned dual branch contrastive features are deeply integrated to minimize classification loss and improve classification results. A large number of experimental results on the ship image dataset FGSC-23 show that the average accuracy of classifying 23 types of fine-grained ship images reaches 83.56%.
2024,46(23): 174-178 收稿日期：2024-2-7
DOI：10.3404/j.issn.1672-7649.2024.23.031
分类号：TP391
基金项目：2022年度江苏省工业和信息产业转型升级专项资金资助项目（CMHI-2022-RDG-004）
作者简介：李英(1985-),男,硕士,工程师,研究方向为电气工程及其自动化
参考文献：
[1] LI B, XIE XY, WEI XX, et al. Ship detection and classification from optical remote sensing images: A survey[J]. Chinese Journal of Aeronautics, 2021, 34(3): 145-163.
[2] 郭璟瑶. 基于视觉传达的模糊舰船图像目标分类检测[J]. 舰船科学技术, 2023, 45(9): 172-175.
GUO J Y. Fuzzy ship image target classification and detection based on visual communication[J]. Ship Science and Technology, 2023, 45(9): 172-175.
[3] 李庆忠, 徐相玉. 基于改进YOLOV3-Tiny的海面船舰目标快速检测[J]. 计算机工程, 2021, 47(10): 283-289.
[4] HUANG L, WANG F, ZHANG Y, et al. Fine-grained ship classification by combining CNN and swin transformer[J]. Remote Sensing, 2022, 14(13): 3087.
[5] WEI X, XIONG Z, and CUI Y. An explainable attention network for fine-grained ship classification using remote-sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, (60): 1-14.
[6] 罗建豪, 吴建鑫. 基于深度卷积特征的细粒度图像分类研究综述[J]. 自动化学报, 2017, 43(8): 1306-1318.
[7] WEI X, SONF Y, AODHA O, et al. Fine-grained image analysis with deep learning: A survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 44(12): 8927-8948.
[8] 黄港, 郑元林, 廖开阳, 等. 互补注意多样性特征融合网络的细粒度分类[J]. 中国图象图形学报, 2023, 28(8): 2420-2431.
[9] ZHANG Y, KANG B, HOOI B, et al. Deep long-tailed learning: A survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023.
[10] 邹承明, 罗莹, 徐晓龙. 基于多特征组合的细粒度图像分类方法[J]. 计算机应用, 2018, 38(7): 1853-1856.
[11] 王阳, 刘立波. 面向细粒度图像分类的双线性残差注意力网络[J]. 激光与光电子学进展, 2020, 57(12): 171-180.
[12] CUI J, LIU S, TIAN Z, et al. Reslt: Residual learning for long-tailed recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 45(3): 3695-3706.
[13] LI R, ZHENG S, DUAN C, et al. Classification of hyperspectral image based on double-branch dual-attention mechanism network[J]. Remote Sensing, 2023, 12(3): 582.
[14] ZHANG S, CHEN C, HU X, et al. Balanced knowledge distillation for long-tailed learning[J]. Neurocomputing, 2023, 527: 36-46.
[15] CHANG D, DING Y, XIE J, et al. The devil is in the channels: Mutual-channel loss for fine-grained image classification[J]. IEEE Transactions on Image Processing, 2020, 29: 4683-4695.
[16] LE-KHAC P H, HEALY G, SMEATON A F. Contrastive representation learning: A framework and review[J]. IEEE Access, 2020, (8): 193907-193934.
[17] CHEN X, HE K. Exploring simple siamese representation learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021.
[18] 张重生, 陈杰, 李岐龙, 等. 深度对比学习综述[J]. 自动化学报, 2023, 49(1): 15-39.
[19] HE K, ZHANG X, REN S, et al, Deep residual learning for image recognition[J]. in Proc. IEEE Conf. Comput. Vis. Pattern Recognit, 2016, 770–778.
[20] DENG J, DONG W, SOCHER R, et al, ImageNet: A large-scale hierarchical image database[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, USA, 2009.
[21] ZHENG H, HU Z, YANG L, et al. Multi-feature collaborative fusion network with deep supervision for SAR ship classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 1-14.
[22] ZHANG X H, LV Y F, YAO L B, et al. A new benchmark and an attribute-guided multilevel feature representation network for finegrained ship classification in optical remote sensing images[J]. IEEE J. Sel. Topics Appl. Earth Observ. Remote Sens., 2020, 13: 1271-1285.
[23] CAO K, WEI C, GAIDON A, et al. Learning imbalanced datasets with label-distribution-aware margin loss[J]. Advances in Neural Information Processing Systems, 2019, 32: 1-12.
[24] WANG X, LIAN L, MIAO Z, et al. Long-tailed recognition by routing diverse distribution-aware experts[C]//International Conference on Learning Representations. 2021.
