Target Detection Algorithm Based on Global Feature Fusion in Parallel Dual Path Backbone

doi:10.3778/j.issn.1673-9418.2312050

Abstract

Abstract: The active downsampling of the backbone of conventional single path architecture often leads to insufficient feature extraction and information loss. At the same time, simply adding or splicing feature pyramids is not conducive to the integration of shallow to deep features. To solve these problems, a target detection algorithm based on global feature fusion in parallel dual path backbone is proposed. Firstly, the dual path architecture backbone is used to extract spatial and semantic information in parallel, and the dual path fusion module is used to promote the mutual complement between feature information. Secondly, the top feature is added to the pyramid pooled multi-scale pool mapping at the same time, and the attention mechanism is used to gather the multi-scale pooled features, so as to further improve the multi-scale detection performance. Then, the global scale information is gathered, which is integrated into different layers of features by using self-attention mechanism, and repeated many times to construct the neck network structure of global feature fusion, which effectively improves the ability of neck network to fuse global context information. Finally, the head adopts Ghost Conv combined with channel shuffling operation to maintain model performance and reduce parameter redundancy. Experiments on KITTI, BDD100K and PASCAL VOC datasets show that the average accuracy of the proposed algorithm is improved by 3.5, 3.4 and 2.7 percentage points compared with the baseline model (YOLOv7-tiny), respectively. Experimental results show that the proposed algorithm improves the detection performance in complex scenes, and has low requirements for computing power and other resources.

Key words: target detection, dual path backbone, pooling attention, global feature fusion neck network, Ghost detection head

摘要： 常规单路径架构主干经过积极的下采样，往往导致特征信息的丢失。同时，仅依靠特征金字塔简单地相加或拼接不利于浅层到深层的特征集成。针对上述问题，提出一种并行双路径主干下全局特征融合的目标检测算法。采用双路径架构主干并行地提取空间与语义信息，并通过双路径融合模块，促进特征信息间的相互补充。顶部特征依次与金字塔池化多尺度池映射相加，利用注意力机制将多尺度池化特征聚集其中，进一步提高多尺度的检测性能。聚集全局尺度信息，利用自注意机制将其融入不同层特征，并重复多次以构建全局特征融合的颈部网络结构，有效提升颈部网络融合全局上下文信息的能力。头部采用Ghost Conv并结合通道混洗操作，维持模型性能的同时减少参数冗余。在KITTI、BDD100K和PASCAL VOC数据集上展开实验，所提算法的平均精度值相较于基线模型（YOLOv7-tiny）分别提高了3.5、3.4和2.7个百分点。实验结果表明，提出的算法提升了复杂场景下的检测性能，而且对算力等资源的要求较低。

关键词: 目标检测, 双路径主干, 池化注意力, 全局特征融合颈部网络, Ghost检测头

QIU Yunfei, XIN Hao. Target Detection Algorithm Based on Global Feature Fusion in Parallel Dual Path Backbone[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(12): 3247-3259.

邱云飞, 辛浩. 并行双路径主干下全局特征融合的目标检测算法[J]. 计算机科学与探索, 2024, 18(12): 3247-3259.

References

[1] SANDLER M, HOWARD A, ZHU M, et al. MobileNetV2: inverted residuals and linear bottlenecks[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2018: 4510-4520.
[2] HOWARD A G, ZHU M, CHEN B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB/OL]. [2023-10-24]. https://arxiv.org/abs/1704.04861.
[3] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings of the 14th European Conference on Computer Vision, Amsterdam, Oct 11-14, 2016. Cham: Springer, 2016: 21-37.
[4] TANG Q, LI J, SHI Z, et al. LightDet: a lightweight and accurate object detection network[C]//Proceedings of the 2020 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway: IEEE, 2020: 2243-2247.
[5] MA N, ZHANG X, ZHENG H T, et al. ShuffleNet V2: practical guidelines for efficient CNN architecture design[C]//Proceedings of the 15th European Conference on Computer Vision. Cham: Springer, 2018: 116-131.
[6] QIN Z, LI Z, ZHANG Z, et al. ThunderNet: towards real-time generic object detection on mobile devices[C]//Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 6718-6727.
[7] 刘鹏, 毕誉轩, 张天翼, 等. 注意力机制优化的全尺寸目标检测方法[J]. 电子测量与仪器学报, 2023, 37(2): 193-203.
LIU P, BI Y X, ZHANG T Y, et al. Full-size object detection method optimized by attention mechanism[J]. Journal of Electronic Measurement and Instrumentation, 2023, 37(2): 193-203.
[8] XIONG Y, LIU H, GUPTA S, et al. MobileDets: searching for object detection architectures for mobile accelerators[C]//Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 3825-3834.
[9] LYU C, ZHANG W, HUANG H, et al. RTMDet: an empirical study of designing real-time object detectors[EB/OL]. [2023-10-24]. https://arxiv.org/abs/2212.07784.
[10] CAI Z, SHEN Q. FalconNet: factorization for the light-weight ConvNets[EB/OL]. [2023-10-24]. https://arxiv.org/abs/2306.06365.
[11] MEHTA S, RASTEGARI M. MobileViT: light-weight, general-purpose, and mobile-friendly vision transformer[EB/OL]. [2023-10-24]. https://arxiv.org/abs/2110.02178.
[12] CHEN Y, DAI X, CHEN D, et al. Mobile-Former: bridging MobileNet and transformer[C]//Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 5270-5279.
[13] LV W, XU S, ZHAO Y, et al. DETRs beat YOLOs on real-time object detection[EB/OL]. [2023-10-24]. https://arxiv.org/abs/2304.08069.
[14] 孙扬, 韩磊, 王程庆, 等. 采用双支路与特征融合网络的路沿分割[J]. 计算机工程与应用, 2023, 59(9): 255-261.
SUN Y, HAN L, WANG C Q, et al. Curb segmentation using dual branch and feature fusion network[J]. Computer Engineering and Applications, 2023, 59(9): 255-261.
[15] 胡学刚, 龚宇, 敬力源. 双路径特征融合编解码结构的高速语义分割[J]. 计算机辅助设计与图形学学报, 2022, 34(12): 1911-1919.
HU X G, GONG Y, JING L Y. High-speed semantic segmentation based on dual-path feature fusion codec structure[J]. Journal of Computer-Aided Design & Computer Graphics,2022, 34(12): 1911-1919.
[16] 聂光涛, 黄华. 光学遥感图像目标检测算法综述[J]. 自动化学报, 2021, 47(8): 1749-1768.
NIE G T, HUANG H. A survey of object detection in optical remote sensing images[J]. Acta Automatica Sinica, 2021, 47(8): 1749-1768.
[17] LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition.Washington: IEEE Computer Society, 2018: 8759-8768.
[18] TAN M, PANG R, LE Q V. EfficientDet: scalable and efficient object detection[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway: IEEE, 2020: 10781-10790.
[19] QIAO S, CHEN L C, YUILLE A. Detectors: detecting objects with recursive feature pyramid and switchable atrous convolution[C]//Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2021: 10213-10224.
[20] YANG G, LEI J, ZHU Z, et al. AFPN: asymptotic feature pyramid network for object detection[EB/OL]. [2023-10-24]. https://arxiv.org/abs/2306.15988.
[21] WANG C Y, BOCHKOVSKIY A, LIAO H Y M. 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.
[22] HAN K, WANG Y, TIAN Q, et al. GhostNet: more features from cheap operations[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 1580-1589.
[23] WANG J, CHEN K, XU R, et al. CARAFE: content-aware reassembly of features[C]//Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2019: 3007-3016.
[24] 吕卫, 梁芷茵, 褚晶辉. 基于改进Anchor-Free模型的交通标志检测算法[J]. 激光与光电子学进展, 2022, 59(24): 167-174.
LV W, LIANG Z Y, CHU J H. Traffic sign detection algorithm based on modified Anchor-Free model[J]. Laser Optoelectronics Progress, 2022, 59(24): 167-174.
[25] 李文举, 储王慧, 崔柳, 等. 结合图采样和图注意力的3D目标检测方法[J]. 计算机工程与应用, 2023, 59(9): 237-244.
LI W J, CHU W H, CUI L, et al. 3D object detection method combining on graph sampling and graph attention[J]. Computer Engineering and Applications, 2023, 59(9): 237-244.
[26] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision. Cham: Springer, 2018: 3-19.
[27] HE K, ZHANG X, REN S, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916.
[28] HONG Y, PAN H, SUN W, et al. Deep dual-resolution networks for real-time and accurate semantic segmentation of road scenes[EB/OL]. [2023-10-24]. https://arxiv.org/abs/2101.06085.
[29] LIU S, HUANG D, WANG Y. Learning spatial fusion for single-shot object detection[EB/OL]. [2023-10-24]. https://arxiv.org/abs/1911.09516.
[30] GEIGER A, LENZ P, URTASUN R. Are we ready for auto-nomous driving? The KITTI vision benchmark suite[C]//Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Recognition. Washington: IEEE Computer Society, 2012: 3354-3361.
[31] YU F, CHEN H, WANG X, et al. BDD100K: a diverse driving dataset for heterogeneous multitask learning[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 2636-2645.
[32] EVERINGHAM M, VAN GOOL L, WILLIAMS C K I, et al. The Pascal visual object classes (VOC) challenge[J]. International Journal of Computer Vision, 2010, 88(2): 303-338.
[33] SELVARAJU R R, COGSWELL M, DAS A, et al. Grad-CAM: visual explanations from deep networks via gradient-based localization[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision. Washington: IEEE Computer Society, 2017: 618-626.
[34] REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems 28, Montreal, Dec 7-12, 2015: 91-99.
[35] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection[EB/OL].[2023-10-24]. https://arxiv.org/abs/2004.10934.
[36] RAO Y, ZHAO W, TANG Y, et al. HorNet: efficient high-order spatial interactions with recursive gated convolutions[C]//Advances in Neural Information Processing Systems 35, New Orleans, Nov 28-Dec 9, 2022: 10353-10366.
[37] LIU Z, MAO H, WU C Y, et al. A convnet for the 2020s[C]//Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2022: 11976-11986.
[38] 周飞, 郭杜杜, 王洋, 等. 基于改进YOLOv8的交通监控车辆检测算法[J]. 计算机工程与应用, 2024, 60(6): 110-120.
ZHOU F, GUO D D, WANG Y, et al. Vehicle detection algorithm based on improved YOLOv8 in traffic surveillance[J]. Computer Engineering and Applications, 2024, 60(6): 110-120.