计算机应用 ›› 2019, Vol. 39 ›› Issue (1): 143-147.DOI: 10.11772/j.issn.1001-9081.2018061194

• 人工智能 • 上一篇    下一篇

深度导向显著性检测算法

赵恒1, 安维胜1, 付为刚2   

  1. 1. 西南交通大学 机械工程学院, 成都 610031;
    2. 中国民用航空飞行学院 航空工程学院, 四川 广汉 618307
  • 收稿日期:2018-06-12 修回日期:2018-07-31 出版日期:2019-01-10 发布日期:2019-01-21
  • 通讯作者: 赵恒
  • 作者简介:赵恒(1993-),男,重庆潼南人,硕士研究生,主要研究方向:图像分割、机器视觉;安维胜(1974-),男,四川资阳人,副教授,博士,主要研究方向:计算机图形学、虚拟现实;付为刚(1984-),男,湖北天门人,副教授,博士,主要研究方向:计算机图形算法。

Saliency detection algorithm of deep guidance

ZHAO Heng1, AN Weisheng1, FU Weigang2   

  1. 1. School of Mechanical Engineering, Southwest Jiaotong University, Chengdu Sichuan 610031, China;
    2. Aviation Engineering Institute, Civil Aviation Flight University of China, Guanghan Sichuan 618307, China
  • Received:2018-06-12 Revised:2018-07-31 Online:2019-01-10 Published:2019-01-21

摘要: 针对目前基于深度卷积神经网络的显著性检测算法存在对复杂场景图像目标检测不完整、背景噪声多的问题,提出一种深度特征导向显著性检测算法。该算法是基于现有底层特征与深度卷积特征融合模型(ELD)的改进,网络模型包含基础特征提取、高层语义特征跨层级引导传递两个部分。首先,根据不同层级卷积特征的差异性,构建跨层级特征联合的高层语义特征引导模型;然后,用改进的网络模型生成初始显著图,利用高层语义特征引导的方式进行显著性聚类传播;最后,用完全联系条件随机场对聚类传播的结果进行优化,使其能够获取更多结构边缘信息和降低噪声并生成完整显著图。在ECSSD上和DUT-ORMON两个数据集上进行实验测试,实验结果表明,所提算法的准确率和召回率(PR)优于ELD模型,其F-measure(F)值分别提升了7.5%和11%,平均绝对误差(MAE)值分别降低了16%和15%,说明了所提算法模型能够在目标识别、模式识别、图像索引等复杂图像场景应用领域得到更加鲁棒的结果。

关键词: 显著性检测, 深度特征, 神经网络, 特征引导, 显著图

Abstract: As current saliency detection algorithms based on deep convolutional network have problems of incomplete target and noisy background detected from complex scene images, a new algorithm of deep feature-oriented saliency detection composed with basic feature extraction and high-level feature which guided cross-level aggregating delivery was proposed. It was based on the improvement of an extant Encoded Low level distance map with Deep features (ELD) model. Firstly, according to the characteristics of convolutional features at different levels, a cross-level feature fusion network model of high-level feature guidance was established. Then, saliency clustering propagation by using high-level feature guidance on initial saliency map that generated by improved neural network was implemented. Finally, final saliency map with more details and less noise was generated by using fully-connected conditional random field after saliency propagation. The experimental results on ECSSD and DUT-ORMON data sets show that, the Precision-Recall (PR) performance of the proposed algorithm is better than ELD algorithms, and F-measure(F) is increased by 7.5% and 11%, respectively, while its Mean Average Errors (MAE) are decreased by 16% and 15%, respectively,which also can obtain more robust results in complex image scene fields of target recognition, pattern recognition, image indexing, and so on.

Key words: saliency detection, deep feature, neural network, feature guidance, salient map

中图分类号: