模式识别与人工智能
2025年4月7日 星期一   首 页     期刊简介     编委会     投稿指南     伦理声明     联系我们                                                                English
模式识别与人工智能  2019, Vol. 32 Issue (11): 1006-1013    DOI: 10.16451/j.cnki.issn1003-6059.201911005
论文与报告 最新目录| 下期目录| 过刊浏览| 高级检索 |
融合对抗学习策略的实例级物体检测算法
覃润楠1, 王睿1
1.北京航空航天大学 仪器科学与光电工程学院 北京 100191
Instance-Level Object Detection Algorithm Fusing Adversarial Learning Strategies
QIN Runnan1, WANG Rui1
1.School of Instrumentation and Optoelectronic Engineering,Beihang University, Beijing 100191

全文: PDF (962 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 针对目前基于深度学习的实例级物体检测算法对受遮挡物体的检测效果较差的问题,文中引入对抗学习的训练策略,提出改进的对抗生成式区域全卷积网络算法(AGR-FCN).以区域全卷积网络(R-FCN)为基准框架,添加为训练样本生成遮挡特征的对抗性遮挡丢弃网络(AMDN).通过R-FCN与AMDN间对抗学习的训练策略,提升R-FCN对遮挡物体的特征学习能力,优化整体实例级物体检测性能.在公共数据库GMU Kitchen和自制数据库BHGI上的实验表明,在复杂多变的非结构化环境中,如随机变化的不同光照、尺度、焦比、视角与姿态、遮挡等条件下,AGR-FCN的平均检测精度较高.
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
覃润楠
王睿
关键词 实例级物体检测对抗学习区域全卷积网络    
Abstract:Existing instance-level object detection algorithms based on deep learning achieve a poor detection effect on occluded objects. To solve the problem, an improved adversarial generated region-based fully convolutional networks(AGR-FCN) with the training strategy of adversarial learning is proposed. The original fully convolutional networks(R-FCN) is regarded as a fiducial frame, and adversarial mask dropout network(AMDN) is designed based on the trained R-FCN to generate occlusion features for training samples. Through the training strategy of adversarial learning between R-FCN and AMDN, the learning ability of R-FCN to the features of occluded objects is improved, and its overall instance-level object detection performance is optimized. Experiments on GMU Kitchen dataset and BHGI dataset show that AGR-FCN algorithm achieves good detection accuracy in complex and changeable unstructured environments, such as randomly varying illumination, scale, focal ratio, angle and attitude and occlusion.
Key wordsInstance-level Object Detection    Adversarial Learning    Region-Based Fully Convolutional Networks   
收稿日期: 2019-06-05     
ZTFLH: TP 391.4  
基金资助:国家自然科学基金项目(No.61673039)资助
通讯作者: 王 睿,博士,副教授,主要研究方向为机器视觉、模式识别和跟踪、光学传感、图像处理等.E-mail:wangr@buaa.edu.cn.   
作者简介: 覃润楠,硕士研究生,主要研究方向为机器视觉、深度学习.E-mail:QinRunnan@buaa.edu.cn.
引用本文:   
覃润楠, 王睿. 融合对抗学习策略的实例级物体检测算法[J]. 模式识别与人工智能, 2019, 32(11): 1006-1013. QIN Runnan, WANG Rui. Instance-Level Object Detection Algorithm Fusing Adversarial Learning Strategies. , 2019, 32(11): 1006-1013.
链接本文:  
http://manu46.magtech.com.cn/Jweb_prai/CN/10.16451/j.cnki.issn1003-6059.201911005      或     http://manu46.magtech.com.cn/Jweb_prai/CN/Y2019/V32/I11/1006
版权所有 © 《模式识别与人工智能》编辑部
地址:安微省合肥市蜀山湖路350号 电话:0551-65591176 传真:0551-65591176 Email:bjb@iim.ac.cn
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn