基于预分割和回归的深度学习目标检测

潘蓉; 孙伟

doi:10.3788/OPE.20172513.0221

您当前的位置：

首页 >

文章列表页 >

基于预分割和回归的深度学习目标检测

信息科学 | 更新时间：2020-08-13

- 基于预分割和回归的深度学习目标检测
- Deep learning target detection based on pre-segmentation and regression
- 光学精密工程 2017年25卷第10s期页码：221-227
- 作者机构：
  
  西安电子科技大学空间科学与技术学院, 陕西西安 710071
- 作者简介：
  
  [ "孙伟(1980-),男,安徽砀山人。主要研究方向为高性能视觉信息计算及嵌入式系统设计。E-mail:wsun@xidian.edu.cn" ]
  [ "潘蓉(1992-),女,山西运城人,研究生,研究方向为数字图像处理。E-mail:1621980248@qq.com" ]
- 基金信息：
  
  国家自然科学基金青年基金资助项目（No.61201290），国家自然科学基金面上资助项目（No.61671356），中央高校基本科研业务资助项目（No.JB161301
- DOI：10.3788/OPE.20172513.0221
  中图分类号： TP752.1
- 收稿日期：2017-06-01，
  
  修回日期：2017-07-04，
  
  纸质出版日期：2017-11-25
- 稿件说明：
移动端阅览
潘蓉, 孙伟,. 基于预分割和回归的深度学习目标检测[J]. 光学精密工程, 2017,25(10s): 221-227

PAN Rong, SUN Wei,. Deep learning target detection based on pre-segmentation and regression[J]. Editorial Office of Optics and Precision Engineering, 2017,25(10s): 221-227
潘蓉, 孙伟,. 基于预分割和回归的深度学习目标检测[J]. 光学精密工程, 2017,25(10s): 221-227 DOI： 10.3788/OPE.20172513.0221.

PAN Rong, SUN Wei,. Deep learning target detection based on pre-segmentation and regression[J]. Editorial Office of Optics and Precision Engineering, 2017,25(10s): 221-227 DOI： 10.3788/OPE.20172513.0221.

摘要

针对高分辨率图像中的小目标检测检测难的问题，结合基于候选区域的目标检测方法中的感兴趣区域提取策略和基于回归的目标检测算法中的回归策略，提出了基于预分割和回归的深度学习目标检测算法。因此使用四叉树对原始图像兴趣目标提取，使用基于回归的目标检测方法对感兴趣区域的目标进行细致的再定位和分类。与传统的Fast-RCNN方法和YOLO系列的基于回归的深度学习方法相比，基于四叉树的深度学习的目标检测算法在精度和速度上有明显优势。经过实验结果分析表明，与Fast-RCNN相比，Quad-ssd算法在目标检测时精度提高了6.5%，达到了74.9%，检测速度大幅提高，达到45帧每秒，完全满足实时性的要求。

Abstract

Aiming at the problem of difficult small target detection in high resolution images

combined with region-of-interest (ROI) extraction strategy in target detection method based on candidate region and regression strategy in target detection algorithm based on regression

deep learning target detection algorithm based on pre-segmentation and regression (Quad-ssd) was proposed. As fast-RCNN series implement image location and classification separately

small targets could be detected but detection time was too long. YOLO series method used regression method to implement classification and location for targets in images at the same time. As only high-level features were used

detection accuracy for small target was not enough. Therefore

quad tree was used to extract interest target of original images

and target detection method based on regression was used to implement detailed relocation and classification for targets in interested region. Compared with traditional Fast-RCNN method and deep learning method based on regression of YOLO series

target detection algorithm of deep learning based on quad tree has obvious advantages in accuracy and speed. The experimental results show that compared with Fast-RCNN

accuracy of Quad-ssd algorithm is improved by 6.5% and reaches 74.9% at the time of target detection. The detection speed is improved greatly; reaching 45 f/s

and can satisfyrequirements of timeliness.

关键词

Keywords

references

宋燕星, 袁峰, 丁振良,等. 使用形态Haar小波法检测目标感兴趣区域[J]. 光学精密工程, 2009, 17(7):1752-1758.

TAIGMAN Y, YANG M, RANZATO M, et al.. DeepFace:Closing the Gap to Human-Level Performance in Face Verification[C]. IEEE Conference on Computer Vision and Pattern Recognition. IEEE Computer Society, 2014:1701-1708.

MA X, GRIMSON W E L. Edge-based rich representation for vehicle classification[C]. Tenth IEEE International Conference on Computer Vision. IEEE, 2005:1185-1192 Vol. 2.

KAZEMI F M, SAMADI S, POORREZA H R, et al.. Vehicle Recognition Using Curvelet Transform and SVM[C]. International Conference on Information Technology. IEEE, 2007:516-521.

FREUND Y, SCHAPIRE R E. A desicion-theoretic generalization of on-line learning and an application to boosting[C]. European Conference on Computational Learning Theory. Springer Berlin Heidelberg, 1995:23-37.

FELZENSZWALB P F, GIRSHICK R B, MCALLESTER D, et al.. Object detection with discriminatively trained part-based models.[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2010, 32(9):1627-45.

SZEGEDY C, TOSHEV A, ERHAN D. Deep Neural Networks for Object Detection.[C]. Advances in Neural Information Processing Systems[S. l.]:NIPS Press, 2013:1673-1675.

SERMANET P, EIGEN D, ZHANG X, et al.. OverFeat:Integrated Recognition, Localization and Detection using Convolutional Networks[J]. Eprint Arxiv, 2013.

GIRSHICK R, DONAHUE J, DARRELL T, et al.. Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation[J]. 2013:580-587.

UIJLINGS J R, SANDE K E, GEVERS T, et al.. Selective Search for Object Recognition[J]. International Journal of Computer Vision, 2013, 104(2):154-171.

HE K, ZHANG X, REN S, et al.. Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 37(9):1904.

GIRSHICK R. Fast R-CNN[J]. Computer Science, 2015.

REN S, HE K, GIRSHICK R, et al.. Faster R-CNN:Towards Real-Time Object Detection with Region Proposal Networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2016, 39(6):1137.

REDMON J, DIVVALA S, GIRSHICK R, et al.. You only look once:unified, real-time object detection[J]. IEEE Computer Society, 2015:779-788.

LIU W, ANGUELOV D, ERHAN D, et al.. SSD:Single Shot MultiBox Detector[C]. European Conference on Computer Vision. Springer International Publishing, 2016:21-37.

LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]. Computer Vision and Pattern Recognition. IEEE, 2015:3431-3440.

RUSSAKOVSKY O, DENG J, SU H, et al.. ImageNet Large Scale Visual Recognition Challenge[J]. International Journal of Computer Vision, 2015, 115(3):211-252.

浏览量

527

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

采用多尺度特征融合SSD的遥感图像小目标检测

采用多尺度特征融合SSD的遥感图像小目标识别

基于多分支残差注意力网络的水下图像增强

基于策略梯度和伪孪生网络的异源图像匹配

面向点云分类和分割的形状自适应特征聚合网络