基于物体间关系和场景信息的目标检测  

Object detection based on relationship between objectsand scene-level information

作　　者：赵维[1] 张燕妮 ZHAO Wei;ZHANG Yan-ni(Department of Information Engineering,Jilin Police College,Changchun 130117,China;School of Information Science and Technology,Northeast Normal University,Changchun 130117,China)

机构地区：[1]吉林警察学院信息工程系,吉林长春130117 [2]东北师范大学信息科学与技术学院,吉林长春130117

出　　处：《东北师大学报（自然科学版）》2020年第2期80-85,共6页Journal of Northeast Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61672150);吉林省科技发展计划项目(20180201089GX,20190201305JC).

摘　　要：提出了一种基于物体间关系和全局场景信息的目标检测方法.该方法利用物体的几何位置信息建立物体间的关系,并将图像的全局场景信息作为目标检测的先验知识.在2个常用的目标检测数据库上测试了方法性能.实验结果表明,所提出的方法不仅具有较好的目标检测准确率,而且具有较高的计算效率.Object detection is an important research topic in the field of computer vision and pattern recognition.In recent years,researchers have proposed many object detection methods,but most of them only consider the local information of the image,while ignoring the interrelationship between objects and the global contextual information in the image.In this paper,we proposed an object detection algorithm based on the relationship between objects and scene-level information which establishes the relationship between objects by the geometric position information of objects and uses the scene-level information of the entire image as a priori knowledge.Our experimental results show that introducing the relationship between objects and the scene-level information into the pipeline of object detection can improve the detection results and speed.

关 键 词：目标检测 物体关系 全局场景信息 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]