渐进式特征增强的弱监督显著性目标检测

doi:10.19678/j.issn.1000-3428.0068406

计算机工程 ›› 2024, Vol. 50 ›› Issue (12): 233-244. doi: 10.19678/j.issn.1000-3428.0068406

渐进式特征增强的弱监督显著性目标检测

李沼洁¹, 朱恒亮¹^,², 毛国君¹^,²^,*(), 杨鑫¹

1. 福建理工大学计算机科学与数学学院, 福建福州 350118
2. 福建省大数据挖掘与应用技术重点实验室, 福建福州 350118

收稿日期:2023-09-17 出版日期:2024-12-15 发布日期:2024-02-21
通讯作者: 毛国君
基金资助:
国家重点研发计划(2019YFD0900905); 国家自然科学基金(61773415); 福建省自然科学基金(2023J01348); 福建理工大学科技项目(GY-Z220205)

Progressively Feature-Enhanced Weakly Supervised for Salient Object Detection

LI Zhaojie¹, ZHU Hengliang¹^,², MAO Guojun¹^,²^,*(), YANG Xin¹

1. College of Computer Science and Mathematics, Fujian University of Technology, Fuzhou 350118, Fujian, China
2. Fujian Provincial Key Laboratory of Big Data Mining and Applications, Fuzhou 350118, Fujian, China

Received:2023-09-17 Online:2024-12-15 Published:2024-02-21
Contact: MAO Guojun

摘要/Abstract

摘要：

针对多数弱监督显著性检测方法在复杂场景下容易出现目标结构缺损、边界粗糙等问题, 提出一种渐进式特征增强的弱监督显著性检测算法。首先针对显著目标结构不完整问题, 设计一种渐进式特征增强机制, 主要包括双流语义增强模块和层次化自适应特征聚合模块, 通过复用这种机制可以捕获更丰富的图像特征; 其次为获取清晰完整的目标边缘, 提出边缘引导模块, 可以生成高质量的显著目标边缘图; 最后将得到的边缘对显著区域预测网络进行指导, 以生成结构完整且边界平滑的检测结果。在5个公开数据集上的实验结果表明, 相比经典的WSSA算法, 该算法在PASCAL-S数据集上平均绝对误差(MAE)降低了21.32%, F-measure值提高了6.27%, 优于大多数先进的弱监督显著性目标检测算法。

关键词: 弱监督, 显著性目标检测, 渐进式, 特征聚合, 边缘引导

Abstract:

For the weakly supervised object detection methods, it is easy to have problems such as object structure defect and rough boundary in complex scenarios. To overcome this problem, a progressively feature-enhanced, weakly supervised network for salient object detection has been proposed. First, this model uses a full convolutional neural network (ResNet-50) as the main stem for feature extraction, which helps the model learn the intrinsic features of the salient regions. To address the problem of the incomplete structure of salient objects, a Progressive Feature-Enhanced Mechanism (PFEM) is designed, which mainly includes a dual-stream semantic enhancement module and a hierarchical adaptive feature aggregation module. By reusing the progressive feature-enhancement mechanism, richer image features can be captured. Moreover, an Edge-Guided Module (EGM) is proposed to exploit the salient object edge structure. EGM can effectively aggregate the boundary formation of features and capture high-quality salient edge maps. Finally, salient edge maps are used to generate detection results with a complete structure and clear boundaries. Experimental results on five public datasets show that, compared with the classical WSSA algorithm, the Mean Absolute Error (MAE) on the PASCAL-S dataset is reduced by 21.32%, and the F-measure is increased by 6.27%, which is better than most advanced weakly supervised significance target detection algorithms.

Key words: weakly supervised, Salient Object Detection(SOD), progressive, feature fusion, edge guided

李沼洁, 朱恒亮, 毛国君, 杨鑫. 渐进式特征增强的弱监督显著性目标检测[J]. 计算机工程, 2024, 50(12): 233-244.

LI Zhaojie, ZHU Hengliang, MAO Guojun, YANG Xin. Progressively Feature-Enhanced Weakly Supervised for Salient Object Detection[J]. Computer Engineering, 2024, 50(12): 233-244.

https://www.ecice06.com/CN/Y2024/V50/I12/233

图/表 14

图1 结构不完整和边界粗糙图像

Fig.1 Images of incomplete structure and rough borders

图2 PFENet网络的整体架构

Fig.2 The overall architecture of PFENet network

图3 渐进式特征融合增强效果

Fig.3 The effect of progressive feature fusion enhancement

图4 双流语义增强模块

Fig.4 Dual-stream semantic enhancement module

图5 层次化自适应特征聚合模块

Fig.5 Hierarchical adaptive feature aggregation module

图6 边缘引导模块

Fig.6 Edge-guided module

图7 不同算法在4个SOD数据集上的F-measure曲线

Fig.7 F-measure curves of different algorithms on 4 SOD datasets

图8 本文算法与其他算法的可视化对比

Fig.8 Visual comparison between the proposed algorithm and other algorithm

图9 失败案例的图示

Fig.9 Illustration of the failed examples

参考文献 42

1	SUN F D , LI W H . Saliency guided deep network for weakly-supervised image segmentation. Pattern Recognition Letters, 2019, 120, 62- 68. doi: 10.1016/j.patrec.2019.01.009
2	HE J F, FENG J Y, LIU X L, et al. Mobile product search with bag of Hash bits and boundary reranking[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2012: 3005-3012.
3	HONG S, YOU T, KWAK S, et al. Online tracking by learning discriminative saliency map with convolutional neural network[EB/OL]. [2023-08-10]. https://arxiv.org/pdf/1502.06796.
4	RUTISHAUSER U, WALTHER D, KOCH C, et al. Is bottom-up attention useful for object recognition? [C]//Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2015: 458-467.
5	XIE C X, XIA C Q, MA M C, et al. Pyramid grafting network for one-stage high resolution saliency detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2022: 11707-11716. URL
6	TIAN X, XU K, YANG X, et al. Bi-directional object-context prioritization learning for saliency ranking[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C, USA: IEEE Press, 2022: 5872-5881. URL
7	YU S Y, XIAO J M, ZHANG B F, et al. Democracy does matter: comprehensive feature mining for co-salient object detection[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D. C., USA: IEEE Press, 2022: 979-988. URL
8	YU S Y , ZHANG B F , XIAO J M , et al. Structure-consistent weakly supervised salient object detection with local saliency coherence. Artificial Intelligence, 2021, 35 (4): 3234- 3242.
9	ZHANG J, YU X, LI A X, et al. Weakly-supervised salient object detection via scribble annotations[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, USA: IEEE Press, 2020: 12543-12552. URL
10	LAI B S, GONG X J. Saliency guided end-to-end learning for weakly supervised object detection[EB/OL]. [2023-08-10]. https://arxiv.org/pdf/1412.26980. URL
11	ZHAO T, WU X Q. Image saliency detection with low-level features enhancement[C]//Proceedings of Chinese Conference on Pattern Recognition and Computer Vision. Berlin, Germany: Springer, 2018: 408-419. URL
12	LI H Y , CHEN J , LU H C , et al. CNN for saliency detection with low-level feature integration. Neurocomputing, 2017, 226, 212- 220. doi: 10.1016/j.neucom.2016.11.056
13	TIAN X. Weakly-supervised salient instance detection (supplementary materials)[EB/OL]. [2023-08-10]. https://arxiv.org/abs/2009.13898.
14	何亚茹, 葛洪伟. 视觉显著区域和主动轮廓结合的图像分割算法. 计算机科学与探索, 2022, 16 (5): 1155- 1168. doi: 10.3778/j.issn.1673-9418.2011043
	HE Y R , GE H W . Image segmentation algorithm combining visual salient regions and active contour. Journal of Frontiers of Computer Science and Technology, 2022, 16 (5): 1155- 1168. doi: 10.3778/j.issn.1673-9418.2011043
15	李军侠, 王星驰, 殷梓, 等. 边缘深度挖掘的弱监督显著性目标检测. 计算机工程, 2022, 48 (3): 1- 12.
	LI J X , WANG X C , YIN Z , et al. Weak supervised salience object detection in deep edge mining. Computer En

选择文件类型/文献管理软件名称

选择包含的内容

渐进式特征增强的弱监督显著性目标检测

Progressively Feature-Enhanced Weakly Supervised for Salient Object Detection

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 42

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

渐进式特征增强的弱监督显著性目标检测

Progressively Feature-Enhanced Weakly Supervised for Salient Object Detection

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 42