Lightweight underwater small object detection based on graph Transformer and RT-DETR

doi:10.11772/j.issn.1001-9081.2025050565

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (5): 1586-1595.DOI: 10.11772/j.issn.1001-9081.2025050565

• Multimedia computing and computer simulation • Previous Articles

Lightweight underwater small object detection based on graph Transformer and RT-DETR

Minqi WU¹, Yuanhua YANG¹(), Hang LI¹, Yaqin HU², Zhihao TANG¹, Teng MEI¹

^1.School of Artificial Intelligence and Computer Science，Hubei Normal University，Huangshi Hubei 435002，China
^2.Wuhan Beike Instruments Company Limited，Wuhan Hubei 430010，China

Received:2025-05-26 Revised:2025-10-13 Accepted:2025-10-20 Online:2025-10-29 Published:2026-05-10
Contact: Yuanhua YANG
About author:WU Minqi， born in 1999， M. S. candidate. His research interests include computer vision， object detection， graph neural networks.
LI Hang， born in 1999， M. S. candidate. His research interests include artificial intelligence.
HU Yaqin， born in 1999， engineer. Her research interests include software engineering.
TANG Zhihao， born in 2000， M. S. candidate. His research interests include deep learning， object detection.
MEI Teng， born in 1991， M. S. candidate. His research interests include computer vision， 3D image processing.
Supported by:
Natural Science Foundation of Hubei Province(2022CFC013);2025 “Graduate Innovation and Research Project” of Hubei Normal University(2025Y127)

基于图Transformer和RT-DETR的轻量化水下小目标检测

吴闵奇¹, 杨元华¹(), 李航¹, 胡雅琴², 汤智豪¹, 梅腾¹

^1.湖北师范大学人工智能与计算机学院，湖北黄石 435002
^2.武汉贝科仪器有限公司，武汉 430010

通讯作者: 杨元华
作者简介:吴闵奇（1999—），男，湖北武汉人，硕士研究生，CCF会员，主要研究方向：计算机视觉、目标检测、图神经网络；
李航（1999—），男，四川成都人，硕士研究生，主要研究方向：人工智能；
胡雅琴（1999—），女，湖北黄石人，工程师，主要研究方向：软件工程；
汤智豪（2000—），男，湖北十堰人，硕士研究生，主要研究方向：深度学习、目标检测；
梅腾（1991—），男，湖北武汉人，硕士研究生，主要研究方向：计算机视觉、三维图像处理。
基金资助:
湖北省自然科学基金资助项目(2022CFC013);2025年湖北师范大学“研究生创新科研”立项建设项目(2025Y127)

Abstract

Abstract:

Existing underwater small object detection methods are primarily based on deep learning algorithms， which face challenges in balancing lightweight design and detection accuracy， so that they unable to meet the requirements of real-time and resource-constrained platforms. Therefore， Graph-DETR， a lightweight underwater small object detection model based on RT-DETR （Real-Time DEtection TRansformer） and a graph Transformer， was proposed. The model used a lightweight MobileNetV4 backbone improved with the Large Separable Kernel Attention mechanism （LSKAttention） and the Context-Mixing dynamic convolutional block （CM block） to enhance feature extraction efficiency and reduce model complexity. Additionally， a hierarchical Graph Transformer Feature Pyramid Network （GTFPN） was proposed to strengthen multi-scale feature fusion， and the hybrid encoder was optimized via Wavelet Transform Convolution （WTConv）， Adaptive downsampling （Adown）， and path pruning， thereby achieving convolutional receptive field expansion of the CNN-based Cross-scale Feature Fusion （CCFF） module with low parameterization. Experimental results on the underwater public dataset URPC2020 show that， compared to RT-DETR， Graph-DETR reduces the parameters by 66.9% and the reasoning latency by 6.8 ms， achieving a mean Average Precision （mAP） of 53.2% and an Average Precision of 86.8% at an IoU threshold of 0.5 （AP@0.5）； on URPC2021， it has 81.3% recall， 54.1% mAP， 87.6% AP@0.5 with only 10.5 ms latency， outperforming the existing methods. Graph-DETR exhibits excellent performance in underwater small object detection and is practical for deployment on resource-constrained underwater platforms.

Key words: underwater small object detection, graph Transformer, RT-DETR (Real-Time DEtection TRansformer), lightweight, multi-scale feature fusion

摘要：

现有的水下小目标检测方法多基于深度学习算法，难以兼顾轻量化与检测精度，无法满足实时性和资源受限水下平台的需求。因此，提出一种基于RT-DETR（Real-Time DEtection TRansformer）和图Transformer的轻量化水下小目标检测模型Graph-DETR。该模型采用大型可分离核注意力机制（LSKAttention）和上下文混合的动态卷积块（CM block）改进的轻量级MobileNetV4作为主干网络，以提升特征提取效率，降低模型复杂度。同时，提出层次化图Transformer金字塔网络（GTFPN）增强多尺度特征融合，并通过小波变换卷积（WTConv）、自适应下采样（Adown）和路径剪枝改进混合编码器，低参数化扩展基于CNN的跨尺度特征融合（CCFF）模块的卷积感受野。在水下公开数据集URPC2020上的实验结果表明，Graph-DETR的参数量比RT-DETR减少66.9%，推理延迟缩短6.8 ms，检测平均精度均值（mAP）达到53.2%，交并比（IoU）阈值为0.5时的平均精度（AP@0.5）为86.8%；在URPC2021上，Graph-DETR的召回率达到81.3%，mAP为54.1%，AP@0.5为87.6%，均超越了对比方法，且推理延迟仅10.5 ms。Graph-DETR展现出优越的水下小目标检测性能，具备实际部署在计算资源受限水下平台的应用前景。

关键词: 水下小目标检测, 图Transformer, RT-DETR, 轻量化, 多尺度特征融合

CLC Number:

TP391.4

Minqi WU, Yuanhua YANG, Hang LI, Yaqin HU, Zhihao TANG, Teng MEI. Lightweight underwater small object detection based on graph Transformer and RT-DETR[J]. Journal of Computer Applications, 2026, 46(5): 1586-1595.

吴闵奇, 杨元华, 李航, 胡雅琴, 汤智豪, 梅腾. 基于图Transformer和RT-DETR的轻量化水下小目标检测[J]. 《计算机应用》唯一官方网站, 2026, 46(5): 1586-1595.

Figures/Tables 13

References 36

[1]	CHEN Y， NIU H， CHEN H， et al. A review of underwater target recognition based on deep learning［J］. Journal of Physics： Conference Series， 2021， 1881（4）： No.042031.
[2]	宋巍，倪舟，梁纪辰，等. 结合目标特征增强与语义-位置路径聚合的水下目标检测［J］. 计算机工程与应用， 2025， 61（15）： 93-110.
	SONG W， NI Z， LIANG J C， et al. Underwater object detection combining target feature enhancement and semantic-location path aggregation［J］. Computer Engineering and Applications， 2025， 61（15）： 93-110.
[3]	HE K， GKIOXARI G， DOLLÁR P， et al. Mask R-CNN［C］// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway： IEEE， 2017： 2980-2988.
[4]	WANG C Y， YEH I H， MARK LIAO H Y. YOLOv9： learning what you want to learn using programmable gradient information［C］// Proceedings of the 2024 European Conference on Computer Vision， LNCS 15089. Cham： Springer， 2025： 1-21.
[5]	Ultralytics. Ultralytics YOLO11 in PyTorch［EB/OL］. ［2025-02-10］. .
[6]	ZHAO X， WANG Z， DENG Z， et al. G-Net： an efficient convolutional network for underwater object detection［J］. Journal of Marine Science and Engineering， 2024， 12（1）： No.116.
[7]	李慧，贾炳志，王晨曦，等. 基于Swin Transformer的生成对抗网络水下图像增强模型［J］. 计算机应用， 2025， 45（5）： 1439-1446.
	LI H， JIA B Z， WANG C X， et al. Generative adversarial network underwater image enhancement model based on Swin Transformer［J］. Journal of Computer Applications， 2025， 45（5）： 1439-1446.
[8]	陈星星，李生林，周香伶. 改进RT-DETR的水下目标检测算法［J］. 激光与光电子学进展， 2025， 62（24）： No.2415001.
	CHEN X X， LI S L， ZHOU X L. Improved underwater object detection algorithm based on improved RT-DETR［J］. Laser and Optoelectronics Progress， 2025， 62（24）： No.2415001.
[9]	陈辉，虞永杰.LGM-YOLOv11：融合多尺度注意力机制的水下目标检测模型［J］.计算机工程与应用，2025，61（23）：248-263.
	CHEN H， YU Y J. LGM-YOLOv11： underwater object detection model fusing multi-scale attention mechanism［J］. Computer Engineering and Applications， 2025， 61（23）： 248-263.
[10]	LIU Z， ZHUANG Y， JIA P， et al. A novel underwater image enhancement algorithm and an improved underwater biological detection pipeline［J］. Journal of Marine Science and Engineering， 2022， 10（9）： No.1204.
[11]	FU C， FAN X， XIAO J， et al. Learning heavily-degraded prior for underwater object detection［J］. IEEE Transactions on Circuits and Systems for Video Technology， 2023， 33（11）： 6887-6896.
[12]	ZHOU J， HE Z， LAM K M， et al. AMSP-UOD： when vortex convolution and stochastic perturbation meet underwater object detection［C］// Proceedings of the 38th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2024： 7659-7667.
[13]	CHEN L， YANG Y， WANG Z， et al. Underwater target detection lightweight algorithm based on multi-scale feature fusion［J］. Journal of Marine Science and Engineering， 2023， 11（2）： No.320.
[14]	ZHAI Z L， NIU N W J， FENG B M， et al. An improved YOLOv8 model enhanced with detail and global features for underwater object detection［J］. Physica Scripta， 2024， 99（9）： No.096008.
[15]	QU S， CUI C， DUAN J， et al. Underwater small target detection under YOLOv8-LA model［J］. Scientific Reports， 2024， 14： No.16108.
[16]	ZHAO Y， LV W， XU S， et al. DETRS beat YOLOs on real-time object detection［C］// Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2024： 16965-16974.
[17]	LOU M， YU Y. OverLoCK： an overview-first-look-closely-next ConvNet with context-mixing dynamic kernels［C］// Proceedings of the 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2025： 128-138.
[18]	QIN D， LEICHNER C， DELAKIS M， et al. MobileNetV4： universal models for the mobile ecosystem［C］// Proceedings of the 2024 European Conference on Computer Vision， LNCS 15098. Cham： Springer， 2025： 78-96.
[19]	LAU K W， PO L M， REHMAN Y A UR. Large separable kernel attention： rethinking the large kernel attention design in CNN［J］. Expert Systems with Applications， 2024， 236： No.121352.
[20]	FINDER S E， AMOYAL R， TREISTER E， et al. Wavelet convolutions for large receptive fields［C］// Proceedings of the 2024 European Conference on Computer Vision， LNCS 15112. Cham： Springer， 2025： 363-380.
[21]	DING X， ZHANG X， MA N， et al. RepVGG： making VGG-style ConvNets great again［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 13728-13737.
[22]	HE K， ZHANG X， REN S， et al. Deep residual learning for image recognition［C］// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2016： 770-778.
[23]	LIU S， QI L， QIN H， et al. Path aggregation network for instance segmentation［C］// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2018： 8759-8768.
[24]	HAN K， WANG Y， GUO J， et al. Vision GNN： an image is worth graph of nodes［C］// Proceedings of the 36th International Conference on Neural Information Processing Systems. Red Hook： Curran Associates Inc.， 2022： 8291-8303.
[25]	LI G， MÜLLER M， THABET A， et al. DeepGCNs： can GCNs go as deep as CNNs？［C］// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2019： 9266-9275.
[26]	VELIČKOVIĆ P， CUCURNLL G， CASANOVA A， et al. Graph attention networks［EB/OL］. ［2025-01-18］..
[27]	YANG J， LIU Z， XIAO S， et al. GraphFormers： GNN-nested transformers for representation learning on textual graph［C］// Proceedings of the 35th International Conference on Neural Information Processing Systems. Red Hook： Curran Associates Inc.， 2021： 28798-28810.
[28]	GE Z， LIU S， WANG F， et al. YOLOX： exceeding YOLO Series in 2021［EB/OL］. ［2025-03-01］..
[29]	LYU R. NanoDet-Plus： super fast and high accuracy lightweight anchor-free object detection model［EB/OL］. ［2025-02-09］..
[30]	JIANG X， ZHUANG X， CHEN J， et al. YOLOv8-MU： an improved YOLOv8 underwater detector based on a large kernel block and a multi-branch reparameterization module［J］. Sensors， 2024， 24（9）： No.2905.
[31]	WANG C Y， BOCHKOVSKIY A， MARK LIAO H Y. YOLOv7： trainable bag-of-freebies sets new state-of-the-art for real-time object detectors［C］// Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2023： 7464-7475.
[32]	QU S， CUI C， DUAN J， et al. Underwater small target detection based on improved YOLOv7 model［EB/OL］. ［2025-03-10］..
[33]	XIN S， GE H， YUAN H， et al. Lightweight underwater target detection algorithm based on YOLOv7［C］// Proceedings of the 6th International Conference on Natural Language Processing. Piscataway： IEEE， 2024： 387-391.
[34]	WANG Y， SUN Y， LIU Z， et al. Dynamic graph CNN for learning on point clouds［J］. ACM Transactions on Graphics， 2019， 38（5）： No.146.
[35]	XU K， HU W， LESKOVEC J， et al. How powerful are graph neural networks？［EB/OL］. ［2025-05-01］..
[36]	HAMILTON W L， YING Z， LESKOVEC J. Inductive representation learning on large graphs［C］// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook： Curran Associates Inc.， 2017： 1025-1035.）

模型	Params/10⁶	GFLOPs	AP/%				mAP/%	AP@0.5/%	Latency/ms
模型	Params/10⁶	GFLOPs	C1	C2	C3	C4	mAP/%	AP@0.5/%	Latency/ms
YOLO11-S^［5］	9.4	21.5	77.5	72.8	86.2	91.6	48.7	82.0	9.4
YOLO11-M^［5］	20.1	68.0	80.5	76.6	87.4	91.7	50.7	84.1	13.3
RT-DETR-S^［16］	20.0	59.6	81.9	77.1	86.6	88.4	49.4	83.5	10.9
RT-DETR-M^［16］	42.7	130.5	83.0	81.3	87.6	90.5	50.0	85.6	16.4
Mask R-CNN^［3］	44.2	260.1	78.8	73.3	70.1	83.6	38.7	76.5	58.4
G-Net^［6］	78.7	33.4	87.0	59.6	74.0	81.1	—	75.4	7.1
YOLOv8-MU^［30］	5.7	28.7	78.5	71.0	84.5	89.4	—	80.9	—
YOLOv5g-lite^［13］	5.7	19.7	69.9	91.4	64.6	84.1	—	77.5	20.7
AMSP-UOD^［12］	25.0	68.0	67.3	87.5	60.6	77.5	40.1	78.5	17.8
RFTM-50^［11］	75.5	195.0	—	—	—	—	48.2	80.7	112.3
YOLOv8-LA^［15］	2.3	6.1	91.5	82.1	87.1	85.9	—	84.2	4.9
YOLOv8-DGF^［14］	6.0	—	75.9	92.5	83.4	87.5	—	84.8	7.1
YOLOX-Nano^［28］	0.9	1.1	61.5	58.0	67.5	71.1	30.5	64.5	3.0
NanoDet-m^［29］	1.0	1.2	59.5	55.7	65.2	68.8	28.2	62.3	2.7
YOLOv7-tiny^［31］	6.2	5.8	68.6	65.1	74.6	78.2	39.9	71.6	6.2
Graph-DETR	14.1	28.4	83.9	81.7	89.0	92.7	53.2	86.8	9.6

模型	Params/10⁶	GFLOPs	AP/%				mAP/%	AP@0.5/%	Latency/ms
模型	Params/10⁶	GFLOPs	C1	C2	C3	C4	mAP/%	AP@0.5/%	Latency/ms
YOLO11-S^［5］	9.4	21.5	77.5	72.8	86.2	91.6	48.7	82.0	9.4
YOLO11-M^［5］	20.1	68.0	80.5	76.6	87.4	91.7	50.7	84.1	13.3
RT-DETR-S^［16］	20.0	59.6	81.9	77.1	86.6	88.4	49.4	83.5	10.9
RT-DETR-M^［16］	42.7	130.5	83.0	81.3	87.6	90.5	50.0	85.6	16.4
Mask R-CNN^［3］	44.2	260.1	78.8	73.3	70.1	83.6	38.7	76.5	58.4
G-Net^［6］	78.7	33.4	87.0	59.6	74.0	81.1	—	75.4	7.1
YOLOv8-MU^［30］	5.7	28.7	78.5	71.0	84.5	89.4	—	80.9	—
YOLOv5g-lite^［13］	5.7	19.7	69.9	91.4	64.6	84.1	—	77.5	20.7
AMSP-UOD^［12］	25.0	68.0	67.3	87.5	60.6	77.5	40.1	78.5	17.8
RFTM-50^［11］	75.5	195.0	—	—	—	—	48.2	80.7	112.3
YOLOv8-LA^［15］	2.3	6.1	91.5	82.1	87.1	85.9	—	84.2	4.9
YOLOv8-DGF^［14］	6.0	—	75.9	92.5	83.4	87.5	—	84.8	7.1
YOLOX-Nano^［28］	0.9	1.1	61.5	58.0	67.5	71.1	30.5	64.5	3.0
NanoDet-m^［29］	1.0	1.2	59.5	55.7	65.2	68.8	28.2	62.3	2.7
YOLOv7-tiny^［31］	6.2	5.8	68.6	65.1	74.6	78.2	39.9	71.6	6.2
Graph-DETR	14.1	28.4	83.9	81.7	89.0	92.7	53.2	86.8	9.6

模型	Params/10⁶	GFLOPs	AP/%				mAP/%	AP@0.5/%	Latency/ms
模型	Params/10⁶	GFLOPs	C1	C2	C3	C4	mAP/%	AP@0.5/%	Latency/ms
YOLO11-S^［5］	9.4	21.5	73.6	91.1	80.6	88.1	49.6	83.3	10.7
YOLO11-M^［5］	20.1	68.0	74.1	91.4	80.1	89.4	50.4	83.7	14.5
RT-DETR-S^［16］	20.0	59.6	78.4	90.7	81.0	88.1	50.7	84.5	11.3
RT-DETR-M^［16］	42.7	130.5	79.4	91.6	82.5	89.9	51.4	85.8	17.0
Mask R-CNN^［3］	44.2	260.1	53.8	80.2	71.0	81.8	38.9	71.7	58.4
YOLOv8-LA^［15］	2.4	7.5	81.7	91.0	80.7	89.7	50.2	84.7	5.3
YOLOv7-GR^［32］	—	—	81.7	91.0	80.7	89.7	50.8	85.8	26.3
YOLOv7-RC^［33］	18.2	69.4	76.9	91.8	83.0	89.0	—	85.2	10.4
YOLOv5s+SAGHS^［10］	7.0	—	79.0	91.8	82.3	89.3	41.7	75.3	8.0
YOLOV5s+CBAM^［10］	7.0	—	82.7	93.4	82.3	91.0	45.1	79.2	11.0
YOLOX-Nano^［28］	0.9	1.1	62.8	70.9	56.3	67.3	31.9	64.3	3.7
NanoDet-m^［29］	1.0	1.2	56.0	70.3	62.2	66.7	29.6	63.8	3.1
YOLOv7-tiny^［31］	6.2	5.8	65.6	80.5	72.1	76.6	41.8	73.7	7.8
Graph-DETR	14.1	28.4	83.1	92.5	84.0	91.0	54.1	87.6	10.5

模型	Params/10⁶	GFLOPs	AP/%				mAP/%	AP@0.5/%	Latency/ms
模型	Params/10⁶	GFLOPs	C1	C2	C3	C4	mAP/%	AP@0.5/%	Latency/ms
YOLO11-S^［5］	9.4	21.5	73.6	91.1	80.6	88.1	49.6	83.3	10.7
YOLO11-M^［5］	20.1	68.0	74.1	91.4	80.1	89.4	50.4	83.7	14.5
RT-DETR-S^［16］	20.0	59.6	78.4	90.7	81.0	88.1	50.7	84.5	11.3
RT-DETR-M^［16］	42.7	130.5	79.4	91.6	82.5	89.9	51.4	85.8	17.0
Mask R-CNN^［3］	44.2	260.1	53.8	80.2	71.0	81.8	38.9	71.7	58.4
YOLOv8-LA^［15］	2.4	7.5	81.7	91.0	80.7	89.7	50.2	84.7	5.3
YOLOv7-GR^［32］	—	—	81.7	91.0	80.7	89.7	50.8	85.8	26.3
YOLOv7-RC^［33］	18.2	69.4	76.9	91.8	83.0	89.0	—	85.2	10.4
YOLOv5s+SAGHS^［10］	7.0	—	79.0	91.8	82.3	89.3	41.7	75.3	8.0
YOLOV5s+CBAM^［10］	7.0	—	82.7	93.4	82.3	91.0	45.1	79.2	11.0
YOLOX-Nano^［28］	0.9	1.1	62.8	70.9	56.3	67.3	31.9	64.3	3.7
NanoDet-m^［29］	1.0	1.2	56.0	70.3	62.2	66.7	29.6	63.8	3.1
YOLOv7-tiny^［31］	6.2	5.8	65.6	80.5	72.1	76.6	41.8	73.7	7.8
Graph-DETR	14.1	28.4	83.1	92.5	84.0	91.0	54.1	87.6	10.5

Backbone	Params/10⁶	GFLOPs	LSKAttention	CM block	GTFPN	Pruning	WTConv	Adown	mAP/%	AP@0.5/%
ResNet50	42.7	130.5							51.4	85.8
ResNet50	46.8	136.2			✓				53.0	87.4
MobileNetV4	11.0	25.8							45.4	80.3
	10.1	21.2	✓						46.7	82.0
	10.1	20.9		✓					46.3	81.1
	14.2	30.9			✓				50.1	84.7
	9.9	21.3					✓		46.0	82.5
	12.3	27.5						✓	45.3	81.4
	10.3	21.1	✓	✓					47.3	84.1
	12.0	26.3	✓	✓			✓	✓	48.2	85.3
	13.7	27.2	✓	✓	✓	✓	✓		53.9	86.9
	14.1	28.4	✓	✓	✓	✓	✓	✓	54.1	87.6

Lightweight underwater small object detection based on graph Transformer and RT-DETR

基于图Transformer和RT-DETR的轻量化水下小目标检测

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 13

References 36

Related Articles 15

Recommended Articles

Metrics

图卷积算子	Params/10⁶	GFLOPs	mAP/%	AP@0.5/%
EdgeConv	14.3	29.6	54.3	86.0
GIN	14.0	28.4	53.7	85.5
GAT	16.7	30.9	54.2	86.9
GraphSAGE	14.4	28.7	54.1	87.1
Max-Relative	14.1	28.4	54.3	87.6

量化策略	量化精度	模型大小/MB	AP@0.5	Latency/ms
PyTorch	FP32	125.4	87.6	78.1
TensorRT	FP16	64.8	87.1	56.5
TensorRT	INT8	46.7	87.3	47.2

[1]	Changzheng XING, Xin ZHENG, Di JIA, Junfeng LIANG. Improved DeepLabV3+ method based on adaptive attention and nested receptive field [J]. Journal of Computer Applications, 2026, 46(5): 1408-1415.
[2]	Hongrui ZHANG, Weiming FENG, Luxia YANG, Yongjie MA. CSAF-YOLO： improved YOLO11 algorithm for underwater small object detection [J]. Journal of Computer Applications, 2026, 46(5): 1578-1585.
[3]	Xinyi YAN, Linglong ZHU, Yonghong ZHANG. CDC-DETR： multi-scale real-time human-vehicle detection method for complex traffic scenarios [J]. Journal of Computer Applications, 2026, 46(4): 1283-1291.
[4]	Hanqing LIU, Guoming SANG, Yijia ZHANG. Remote sensing image captioning model combining dense multi-scale feature fusion and feature knowledge-enhanced Transformer [J]. Journal of Computer Applications, 2026, 46(3): 741-749.
[5]	Yinshan YU, Xu TANG, Mingjian DING, Wenkai HUANG, Jiawen BI, Guochen TAN. Real-time vehicle detection algorithm based on YOLOv10 [J]. Journal of Computer Applications, 2026, 46(3): 950-958.
[6]	Ping HUANG, Qing LI, Haifeng QIU, Chengsi WANG, Anzi HUANG, Xiang ZHANG. Real-time face blurring method based on head skeleton point detection [J]. Journal of Computer Applications, 2026, 46(2): 596-603.
[7]	Ning CAO, Xin WEN, Yanrong HAO, Rui CAO. Lightweight motor imagery electroencephalogram decoding neural network with multi-domain feature fusion [J]. Journal of Computer Applications, 2026, 46(1): 289-296.
[8]	Yiming LIANG, Jing FAN, Wenze CHAI. Multi-scale feature fusion sentiment classification based on bidirectional cross attention [J]. Journal of Computer Applications, 2025, 45(9): 2773-2782.
[9]	Liang CHEN, Xuan WANG, Kun LEI. Helmet wearing detection algorithm for complex scenarios based on cross-layer multi-scale feature fusion [J]. Journal of Computer Applications, 2025, 45(7): 2333-2341.
[10]	Dehui ZHOU, Jun ZHAO, Jinfeng CHENG. Tiny defect detection algorithm for bearing surface based on RT-DETR [J]. Journal of Computer Applications, 2025, 45(6): 1987-1997.
[11]	Xiang WANG, Qianqian CUI, Xiaoming ZHANG, Jianchao WANG, Zhenzhou WANG, Jialin SONG. Wireless capsule endoscopy image classification model based on improved ConvNeXt [J]. Journal of Computer Applications, 2025, 45(6): 2016-2024.
[12]	Shiyue GUO, Jianwu DANG, Yangping WANG, Jiu YONG. 3D hand pose estimation combining attention mechanism and multi-scale feature fusion [J]. Journal of Computer Applications, 2025, 45(4): 1293-1299.
[13]	Zhongwei ZHANG, Jun WANG, Shudong LIU, Zhiheng WANG. Object detection in remote sensing image based on multi-scale feature fusion and weighted boxes fusion [J]. Journal of Computer Applications, 2025, 45(2): 633-639.
[14]	Haiteng MENG, Xiaole ZHAO, Tianrui LI. Lightweight image super-resolution reconstruction based on asymmetric information distillation network [J]. Journal of Computer Applications, 2025, 45(2): 601-609.
[15]	Songsen YU, Zhifan LIN, Guopeng XUE, Jianyu XU. Lightweight large-format tile defect detection algorithm based on improved YOLOv8 [J]. Journal of Computer Applications, 2025, 45(2): 647-654.