|
|
论文:2017,Vol:35,Issue(1):170-176 |
|
|
引用本文： |
|
|
黄伟, 杨文姬, 曾璟, 曾舒如, 陈光. 基于谱聚类和增量学习的运动目标物体检测算法研究[J]. 西北工业大学学报 |
|
|
Huang Wei, Yang Wenji, Zeng Jing, Zeng Shuru, Chen Guang. A Novel Algorithm of Moving Object Detection via Spectral Clustering and Incremental Learning[J]. Northwestern polytechnical university |
|
|
|
|
|
|
|
| 基于谱聚类和增量学习的运动目标物体检测算法研究 |
|
| 黄伟1, 杨文姬2, 曾璟1, 曾舒如3, 陈光4 |
|
1. 南昌大学 信息工程学院, 江西 南昌 330031;
2. 江西农业大学 软件学院, 江西 南昌 330045;
3. 华中光电技术研究所, 湖北 武汉 430223;
4. 西安通信学院, 陕西 西安 710072 |
| 摘要: |
| 运动目标物体检测是计算机视觉领域的热门研究方向之一。该方向的一些复杂问题，例如：环境光照变化、目标物体部分/全遮挡、目标物体刚性/非刚性形变等，仍极具挑战性，并制约检测算法效果的进一步提高。为此，提出了一种新颖的运动目标物体检测算法。该算法采用了增量学习技术，融合了视频相邻帧在空间和时间上的高相关性，在每个测试帧上都利用其相邻帧的训练数据进行模型的自学习与更新，从而保证了模型在不同环境或复杂背景下能自动调整。为了实现模型学习，还提出并采用了一种新颖的谱聚类技术。该算法通过一个由1 000多帧的视频数据库验证，采用统计学中的方差分析和多重对比等实验手段，综合分析了该算法与其他同类经典算法的效果。通过大量统计分析，结果表明，该新颖检测算法比传统算法在运动目标物体检测的准确性和鲁棒性上都有明显提高。 |
| 关键词:
算法
谱聚类
增量学习
运动目标物体
检测
|
|
| A Novel Algorithm of Moving Object Detection via Spectral Clustering and Incremental Learning |
|
| Huang Wei1, Yang Wenji2, Zeng Jing1, Zeng Shuru3, Chen Guang4 |
|
1. School of Information Engineering, Nanchang University, Nanchang 330031, China;
2. School of Software Engineering, Jiangxi Agriculture University, Nanchang 330045, China;
3. Huazhong Institute of Electro-Opticis, Wuhan 430223, China;
4. Xi'an Communication Institute, Xi'an 710072, China |
| Abstract: |
| Moving object detection receives much research interest in contemporary computer vision studies. It is also widely acknowledged that many problems, including illumination changes, partial or full occlusions, rigid or non-rigid shape transformation, are still challenging and hinder the detection performance from being further improved. In this paper, a novel algorithm of moving object detection is introduced. The incremental learning technique is employed in this algorithm; whose main purpose is to incorporate high spatial correlation within individual frames as well as high temporary correlation between consecutive frames for automatic updating of the detection model. Also, the model learning is realized via spectral clustering. A databased composed of over 1000 video frames is utilized for experimental evaluation. A series of statistical analysis, including ANOVA and post-hoc multiple comparison tests, are implemented to evaluate the new algorithm and other compared methods. It turns out that the novel algorithm can outperform others in terms of detection accuracy and robustness from the statistical perspective. |
| Key words:
algorithms
spectral clustering
incremental learning
moving targets
detection
analysis of variance (ANOVA)
clustering
constrained optimization
pixels
support vector machines
target tracking
|
|
| 收稿日期: 2016-09-19
修回日期:
|
| DOI: |
| 基金项目: 国家自然科学基金（61403182、61363046）资助 |
| 通讯作者: 杨文姬(1984-),江西农业大学博士研究生,主要从事计算机视觉研究。
Email： |
| 作者简介: 黄伟(1983-),南昌大学副教授,主要从事图像处理研究。
|
|
| 相关功能 |
|
|
|
|
| 作者相关文章 |
|
| 黄伟 在本刊中的所有文章 |
| 杨文姬 在本刊中的所有文章 |
| 曾璟 在本刊中的所有文章 |
| 曾舒如 在本刊中的所有文章 |
| 陈光 在本刊中的所有文章 |
|
|
|
|
|
|
|
|
参考文献: |
|
|
[1] Yimaz A, Javed O, Shah M. Object Tracking:A Survey[J]. ACM Computing Surveys, 2006, 38(4):1-45
[2] Luo W, Xing J, Zhang X, Zhao X, Kim T. Multiple Object Tracking:A Liturature Review[EB/OL]. (2015-09-21)[2016-04-19]. http://arxiv.org/abs/1409.7618
[3] Xu C, Tao D, Xu C. Multiview Intact Space Learning[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2015, 37(12):2531-2544
[4] Hong Z, Chen Z, Wang C, Mei X, Prokhorov D, Tao D. Multi-Store Tracker (MUSTer):A Cognitive Psychology Inspired Approach to Object localization[C]//IEEE Conference on Computer Vision and Pattern Recognition, 2015:749-758
[5] Zhang P, Wang L, Huang W, Xie L, Chen G. Multiple Pedestrian Localization Based on Couple-States Markov Chain with Semantic Topic Learning for Video Surveillance[J]. Soft Computing, 2015, 19(1):85-97
[6] Zhuo T, Zhang P, Zhang Y, Huang W, Sahli H. Object Localization Using Reformative Transductive Learning with Sample Variational Correspondence[C]//ACM International Conference on Multimedia, 2014:941-944
[7] Zhuo T, Zhang Y, Zhang P, Huang W, Sahli H. Non-Rigid Target Localization Based on Flow-Cut in Pair-Wise Frames with Online Hough Forests[C]//ACM International Conference on Multimedia, 2013:489-492
[8] Shi J, Malik J. Normalized Cuts and Image Segmentation[J]. IEEE Trans on Pattern Analysis and Machine Intelligence, 2000, 22:888-905
[9] Ng A, Jordan M, Weiss Y. On Spectral Clustering:Analysis and Algorithm[C]//Advance in Neural Information Processing Systems, 2002:64-72
[10] 邓力, 俞栋, 谢磊. 深度学习:方法及应用[M]. 北京:机械工业出版社, 2016:32-64 Deng Li, Yu Dong, Xie Lei. Deep Learning:Methods and Applications[M]. Beijing, China Machine Press, 2016:32-64(in Chinese) |
|
|
|
相关文献: |
|
|
1．张兆林, 李路, 姚如贵, 王伶.基于LMS算法的窄带干扰检测技术[J]. 西北工业大学学报, 2016,34(1): 92-97 |
|
|
2．卢苇, 周韬, 邢薇薇.一种改进的非结构化P2P网络洪泛搜索机制[J]. 西北工业大学学报, 2015,33(2): 342-350 |
|
|
3．黄登山, 王晓楠.基于Canny和Hillbert-Huang变换的改进的SIFT算法研究[J]. 西北工业大学学报, 2015,33(6): 962-965 |
|
|
4．郑炜, 李知隆, 靳如一.基于频率差异积分的故障定位算法研究[J]. 西北工业大学学报, 2013,31(3): 435-439 |
|
|
5．邓正宏, 徐继伟, 周岩, 乔洪海.基于循环平稳检测算法SSCA的FFT模块设计[J]. 西北工业大学学报, 2013,31(5): 706-709 |
|
|
6．王静, 黄建国, 张群飞, 韩晶.噪声空间谱预白化小孔径阵列被动目标检测方法[J]. 西北工业大学学报, 2012,30(3): 422-427 |
|
|
|
|
|
|
