Pub Date : 2011-12-01DOI: 10.1109/IVSURV.2011.6157028
C. Rao, Shuoqi Liu
When used in practical applications, the speed of intelligent visual surveillance algorithms may decline dramatically due to massive data. Thus the computing speed of algorithms can be a crucial factor in the practical applications. In addition to excellent parallel computing capability, a modern GPU also has large bandwidth and powerful floating-point computing capability. These features make GPU an appropriate device for doing general-purpose computing. This paper accelerates Gaussian Mixture Model and HLSIFT (Harris-like Scale Invariant Feature Detector) using CUDA. The former algorithm gets more than 45 times accelerating and the latter one gets more than 35 times accelerating. The acceleration result is impressive.
{"title":"Research of CUDA in intelligent visual surveillance algorithms","authors":"C. Rao, Shuoqi Liu","doi":"10.1109/IVSURV.2011.6157028","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157028","url":null,"abstract":"When used in practical applications, the speed of intelligent visual surveillance algorithms may decline dramatically due to massive data. Thus the computing speed of algorithms can be a crucial factor in the practical applications. In addition to excellent parallel computing capability, a modern GPU also has large bandwidth and powerful floating-point computing capability. These features make GPU an appropriate device for doing general-purpose computing. This paper accelerates Gaussian Mixture Model and HLSIFT (Harris-like Scale Invariant Feature Detector) using CUDA. The former algorithm gets more than 45 times accelerating and the latter one gets more than 35 times accelerating. The acceleration result is impressive.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"303 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128625945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2011-12-01DOI: 10.1109/IVSURV.2011.6157021
Zhaoxiang Zhang, Yuqing Hou, Yunhong Wang, Jie Qin
Traffic flow detection plays an important role in Intelligent Transportation Systems(ITS). Video based traffic flow detection system is the most widely used strategy in ITS. Under this circumstance, we design and implement a video based traffic flow detection system which is called MyTD in this paper. MyTD takes advantages of both shadow removal and optical flow algorithms. Firstly, we introduce the current development of ITS and focus on the video based traffic detection technology, which is the key to ITS. Secondly, a shadow removal algorithm combining information in both RGB and HSV color spaces is proposed. Thirdly, based on the Iterative Pyramidal LK Optical Flow Algorithm, a vehicle tracking function is realized by OpenCV, as well as a Connected Components Labeling function and the vehicle counting function. Finally, MyTD is implemented and tested based on the algorithm presented above. Experimental results show the outstanding performance of our method comparing with traditional optical flow algorithms.
{"title":"A traffic flow detection system combining optical flow and shadow removal","authors":"Zhaoxiang Zhang, Yuqing Hou, Yunhong Wang, Jie Qin","doi":"10.1109/IVSURV.2011.6157021","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157021","url":null,"abstract":"Traffic flow detection plays an important role in Intelligent Transportation Systems(ITS). Video based traffic flow detection system is the most widely used strategy in ITS. Under this circumstance, we design and implement a video based traffic flow detection system which is called MyTD in this paper. MyTD takes advantages of both shadow removal and optical flow algorithms. Firstly, we introduce the current development of ITS and focus on the video based traffic detection technology, which is the key to ITS. Secondly, a shadow removal algorithm combining information in both RGB and HSV color spaces is proposed. Thirdly, based on the Iterative Pyramidal LK Optical Flow Algorithm, a vehicle tracking function is realized by OpenCV, as well as a Connected Components Labeling function and the vehicle counting function. Finally, MyTD is implemented and tested based on the algorithm presented above. Experimental results show the outstanding performance of our method comparing with traditional optical flow algorithms.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130483724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2011-12-01DOI: 10.1109/IVSURV.2011.6157015
Jian Wang, W. Hu, Zhiling Wang, Muhammad Sarfaraz Malik, Zonghai Chen
The video sequence which contains certain human action is considered as a spatio-temporal volume. There exists certain characteristic signature in appropriately selected spatio-temporal slice of the video sequence. By using these discriminative signatures which we call “human action logos”, new approaches are proposed for period detection and action recognition. Algorithm performance is evaluated under eight typical human actions. Preliminary experiments have shown promising results.
{"title":"Logos of human actions","authors":"Jian Wang, W. Hu, Zhiling Wang, Muhammad Sarfaraz Malik, Zonghai Chen","doi":"10.1109/IVSURV.2011.6157015","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157015","url":null,"abstract":"The video sequence which contains certain human action is considered as a spatio-temporal volume. There exists certain characteristic signature in appropriately selected spatio-temporal slice of the video sequence. By using these discriminative signatures which we call “human action logos”, new approaches are proposed for period detection and action recognition. Algorithm performance is evaluated under eight typical human actions. Preliminary experiments have shown promising results.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123318034","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2011-12-01DOI: 10.1109/IVSURV.2011.6157026
Xia Xuan, Liu Huaping, Xu Weiming, Sun Fuchun
Implementation of particle filter visual tracking on DSP platform will suffer from calculation bottleneck. To realize the real-time tracking, this paper uses the incremental histogram calculation algorithm to construct the histogram of color and edge orientation, integrates the histograms for the observation model and optimizes the target tracking algorithm on the DSP. The experiment proves that the algorithm is fast and the robustness of the system.
{"title":"DSP-based incremental histogram calculation and particle filter tracking algorithm and its implementation","authors":"Xia Xuan, Liu Huaping, Xu Weiming, Sun Fuchun","doi":"10.1109/IVSURV.2011.6157026","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157026","url":null,"abstract":"Implementation of particle filter visual tracking on DSP platform will suffer from calculation bottleneck. To realize the real-time tracking, this paper uses the incremental histogram calculation algorithm to construct the histogram of color and edge orientation, integrates the histograms for the observation model and optimizes the target tracking algorithm on the DSP. The experiment proves that the algorithm is fast and the robustness of the system.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114450994","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2011-12-01DOI: 10.1109/IVSURV.2011.6157019
Haibo Sun, Lijun Cao, Yuan-lu Xie, Mingrui Zhao
This paper has studied on method of video synopsis based on maximum motion power, the events of interest were obtained by the method of background modeling and target tracking. Then these events had a time-shifting and recombinant, as well as calculating the maximum motion power to get the best length of video summary, the video summary had been created finally. The formula of video motion power was derived and validated, the problem of video summary had been transformed into the problem of calculating the biggest motion power of original video. This method can maintain the integrity of the original video information of interest while making the minimum length of the summary video.
{"title":"The method of video synopsis based on maximum motion power","authors":"Haibo Sun, Lijun Cao, Yuan-lu Xie, Mingrui Zhao","doi":"10.1109/IVSURV.2011.6157019","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157019","url":null,"abstract":"This paper has studied on method of video synopsis based on maximum motion power, the events of interest were obtained by the method of background modeling and target tracking. Then these events had a time-shifting and recombinant, as well as calculating the maximum motion power to get the best length of video summary, the video summary had been created finally. The formula of video motion power was derived and validated, the problem of video summary had been transformed into the problem of calculating the biggest motion power of original video. This method can maintain the integrity of the original video information of interest while making the minimum length of the summary video.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"116 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123151305","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper describes a novel object-tracking algorithm by classifying the pixels in a search area into “target” and “background” with K-means clustering algorithm. Two improvements are made to the conventional K-means tracker to solve the instability problem that occurs when some background objects show similar colors to the target or the size of the target object changes significantly. The first one is introducing of the depth information as the sixth feature into the original 5D feature space for describing pixels. The second one is to use Mahalanobis distance in order to keep the balance between color and position when evaluating the difference between pixels. EK-means Tracker can track non-rigid object and wired object at video rate. Its effectiveness was confirmed through several comparison experiments.
{"title":"EK-means tracker: A pixel-wise tracking algorithm using kinect","authors":"Yiqiang Qi, Kazumasa Suzuki, Haiyuan Wu, Qian Chen","doi":"10.1109/IVSURV.2011.6157029","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157029","url":null,"abstract":"This paper describes a novel object-tracking algorithm by classifying the pixels in a search area into “target” and “background” with K-means clustering algorithm. Two improvements are made to the conventional K-means tracker to solve the instability problem that occurs when some background objects show similar colors to the target or the size of the target object changes significantly. The first one is introducing of the depth information as the sixth feature into the original 5D feature space for describing pixels. The second one is to use Mahalanobis distance in order to keep the balance between color and position when evaluating the difference between pixels. EK-means Tracker can track non-rigid object and wired object at video rate. Its effectiveness was confirmed through several comparison experiments.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115372613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2011-10-23DOI: 10.1007/978-3-642-31919-8_21
Lei Xu, Yue Zhou, Qingshan Li
{"title":"Region based on object recognition in 3D scenes","authors":"Lei Xu, Yue Zhou, Qingshan Li","doi":"10.1007/978-3-642-31919-8_21","DOIUrl":"https://doi.org/10.1007/978-3-642-31919-8_21","url":null,"abstract":"","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"172 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2011-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125788016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：