《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (6): 1910-1918.DOI: 10.11772/j.issn.1001-9081.2022050706

• 多媒体计算与计算机仿真 • 上一篇    下一篇

基于边界框标注的弱监督显著性目标检测算法

王强1,2, 黄小明1,2, 佟强1,2, 刘秀磊1,2()   

  1. 1.北京信息科技大学 数据科学与情报分析研究所, 北京 100101
    2.北京材料基因工程高精尖创新中心(北京信息科技大学), 北京 100101
  • 收稿日期:2022-05-18 修回日期:2023-01-04 接受日期:2023-01-10 发布日期:2023-06-08 出版日期:2023-06-10
  • 通讯作者: 刘秀磊
  • 作者简介:王强(1996—),男,安徽潜山人,硕士研究生,主要研究方向:机器学习、图像识别
    黄小明(1977—),男,安徽潜山人,副教授,博士,CCF会员,主要研究方向:机器学习、目标检测、语义分割
    佟强(1985—),男(锡伯族),辽宁沈阳人,讲师,博士,CCF会员,主要研究方向:图像识别、计算机视觉、机器学习
    刘秀磊(1981—),男,河南濮阳人,教授,博士,CCF会员,主要研究方向:语义 Web、本体匹配、语义搜索、知识图谱、语义传感器Email:liuxiulei@bistu.edu.cn
  • 基金资助:
    国家重点研发计划项目(2021YFB2600600);北京信息科技大学校级基金资助项目(2121YJPY225);科研机构创新能力建设;北京市教委科研计划项目(KM202011232014)

Weakly supervised salient object detection algorithm based on bounding box annotation

Qiang WANG1,2, Xiaoming HUANG1,2, Qiang TONG1,2, Xiulei LIU1,2()   

  1. 1.Institute of Data Science and Information Analysis,Beijing Information Science and Technology University,Beijing 100101,China
    2.Beijing Advanced Innovation Center for Materials Genome Engineering (Beijing Information Science and Technology University),Beijing 100101,China
  • Received:2022-05-18 Revised:2023-01-04 Accepted:2023-01-10 Online:2023-06-08 Published:2023-06-10
  • Contact: Xiulei LIU
  • About author:WANG Qiang, born in 1996, M. S. candidate. His research interests include machine learning, image recognition.
    TONG Qiang, born in 1985, Ph. D., lecturer. His research interests include image recognition, computer vision, machine learning.
    First author contact:HUNAG Xiaoming, born in 1977, Ph. D., associate professor. His research interests include machine learning, object detection, semantic segmentation.
  • Supported by:
    National Key Research and Development Program of China(2021YFB2600600);Fund of Beijing Information Science and Technology University(2121YJPY225);Innovation Capacity Building of Scientific Research Institutions,Beijing Municipal Education Commission Science and Technology Program(KM202011232014)

摘要:

针对以往的弱监督显著性目标检测算法存在的显著目标定位不准确问题,提出一种基于边界框标注的弱监督显著目标检测算法。所提算法利用图像中所有目标的最小外接矩形框,即边界框,作为监督信息。首先基于边界框标注和GrabCut算法生成初始显著图;然后在此基础上设计了一个缺失修正模块,以得到优化后的显著图;最后结合传统方法和深度学习方法各自的优势,将优化后的显著图作为伪真值,通过神经网络学习一个显著性目标检测模型。在4个公开数据集上与6种无监督、4种弱监督的显著性检测算法进行比较的实验结果显示,所提算法在所有数据集上的最大F度量值(Max-F)和平均绝对误差(MAE)均明显优于对比算法:与同样基于边界框标注的弱监督方法SBB(Saliency Bounding Boxes)相比,所提算法的标注方法更简单,在ECSSD、DUTS-TE、HKU-IS、DUT-OMRON等4个数据集上进行实验,Max-F分别提高了1.82%、4.00%、1.27%和5.33%,MAE分别降低了13.89%、15.07%、8.77%和13.33%。可见,所提算法是一种具有良好检测性能的弱监督显著目标检测算法。

关键词: 弱监督, 边界框标注, 显著图, 伪真值, 显著性目标检测

Abstract:

Aiming at the inaccurate positioning problem of salient object in the previous weakly supervised salient object detection algorithms, a weakly supervised salient object detection algorithm based on bounding box annotation was proposed. In the proposed algorithm, the minimum bounding rectangle boxes, which are the bounding boxes of all objects in the image were adopted as supervision information. Firstly, the initial saliency map was generated based on the bounding box annotation and GrabCut algorithm. Then, a correction module for missing object was designed to obtain the optimized saliency map. Finally, by combining the advantages of the traditional methods and deep learning methods, the optimized saliency map was used as the pseudo ground-truth to learn a salient object detection model through neural network. Comparison of the proposed algorithm and six unsupervised and four weakly supervised saliency detection algorithms was carried on four public datasets. Experimental results show that the proposed algorithm significantly outperforms comparison algorithms in both Max F-measure value (Max-F) and Mean Absolute Error (MAE) on four datasets. Compared with SBB (Sales Bounding Boxes), which is also a weakly supervised method based on boundary box annotation, the annotation method of the proposed algorithm is simpler. Experiments were conducted on four datasets, ECSSD, DUTS-TE, HKU-IS, DUT-OMRON, and the Max-F increased by 1.82%, 4.00%, 1.27% and 5.33% respectively, and the MAE decreased by 13.89%, 15.07%, 8.77% and 13.33%, respectively. It can be seen that the proposed algorithm is a weakly supervised salient object detection algorithm with good detection performance.

Key words: weakly supervised, bounding box annotation, saliency map, pseudo ground-truth, salient object detection

中图分类号: