Pub Date : 2011-12-01DOI: 10.1109/IVSURV.2011.6157010
Meibin Qi, B. Geng, Jianguo Jiang, Tiao Wang
Due to the random spatial distribution and fast motion of rain, removal of rain in video is a more difficult problem. This paper presents a background subtraction based on sample model to remove the rain. First, analyze the properties of rain, then establish the sample model with values randomly take in the spatial neighborhood of each pixel on the first frame, so better to classify detected rain by background subtraction. In addition, the movement of objects will cause the corresponding color pixel brightness values to change significantly, the H component of the HSI color space was applied to reduce the impact of moving objects on rain removal. Experimental results show that this method compared with existing methods can not only eliminate a good rain, but also have faster processing speed.
{"title":"A rain detection and removal method in video image","authors":"Meibin Qi, B. Geng, Jianguo Jiang, Tiao Wang","doi":"10.1109/IVSURV.2011.6157010","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157010","url":null,"abstract":"Due to the random spatial distribution and fast motion of rain, removal of rain in video is a more difficult problem. This paper presents a background subtraction based on sample model to remove the rain. First, analyze the properties of rain, then establish the sample model with values randomly take in the spatial neighborhood of each pixel on the first frame, so better to classify detected rain by background subtraction. In addition, the movement of objects will cause the corresponding color pixel brightness values to change significantly, the H component of the HSI color space was applied to reduce the impact of moving objects on rain removal. Experimental results show that this method compared with existing methods can not only eliminate a good rain, but also have faster processing speed.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"21 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":"121532863","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.2011.6157036
Ya-Li Hou, G. Pang
Crowd segmentation is an important topic in a visual surveillance system. In this paper, crowd segmentation is formulated as a problem to cluster the feature points inside the foreground region with a set of rectangles. Coherent motion of feature points in an individual are fused with appearance cues around the feature points for crowd segmentation, which has improved the segmentation performance. Furthermore, three descriptors are proposed to extract the points with a non-articulated movement. Some results on the CAVIAR dataset have been shown. The results show that coherent motion cue can be used more reliably by considering the points with rigid motion only.
{"title":"Crowd segmentation based on fusion of appearance and motion features","authors":"Ya-Li Hou, G. Pang","doi":"10.1109/IVSURV.2011.6157036","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157036","url":null,"abstract":"Crowd segmentation is an important topic in a visual surveillance system. In this paper, crowd segmentation is formulated as a problem to cluster the feature points inside the foreground region with a set of rectangles. Coherent motion of feature points in an individual are fused with appearance cues around the feature points for crowd segmentation, which has improved the segmentation performance. Furthermore, three descriptors are proposed to extract the points with a non-articulated movement. Some results on the CAVIAR dataset have been shown. The results show that coherent motion cue can be used more reliably by considering the points with rigid motion only.","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":"129677166","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.2011.6157027
Xiaoyi Yu, Han Zhou, Lingyi Wu, Qingfeng Liu
Machine vision systems have been designed for automated monitoring and analysis of social behavior in Drosophila by Herko Dankert. The Ctrax (The Caltech Multiple Fly Tracker) is implemented for tracking the Drosophila's movement. But the machine vision method is so sophisticated that it is hard to use by a researcher who is lack of computer technology knowledge. Likewise, most of the machine vision solutions are poor performance for the real time environment. Our work focuses on developing a high-performance application to track moving Drosophila and generate reliable tracks from the video. A light solution to automatically tracking the Drosophila in video is implemented, based on object motion in different parts of the scene. The results of effect tests show that our solution tackles the questions of performance, usability and accuracy in machine vision for bioresearch.
{"title":"High-performance drosophila movement tracking","authors":"Xiaoyi Yu, Han Zhou, Lingyi Wu, Qingfeng Liu","doi":"10.1109/IVSURV.2011.6157027","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157027","url":null,"abstract":"Machine vision systems have been designed for automated monitoring and analysis of social behavior in Drosophila by Herko Dankert. The Ctrax (The Caltech Multiple Fly Tracker) is implemented for tracking the Drosophila's movement. But the machine vision method is so sophisticated that it is hard to use by a researcher who is lack of computer technology knowledge. Likewise, most of the machine vision solutions are poor performance for the real time environment. Our work focuses on developing a high-performance application to track moving Drosophila and generate reliable tracks from the video. A light solution to automatically tracking the Drosophila in video is implemented, based on object motion in different parts of the scene. The results of effect tests show that our solution tackles the questions of performance, usability and accuracy in machine vision for bioresearch.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"23 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":"134519762","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.2011.6157014
Xin Zhao, Kaiqi Huang, T. Tan
Most visual surveillance and video understanding systems require knowledge of categories of objects in the scene. One of the key challenges is to be able to classify any object in a real-time procedure in spite of changes in the scene over time and the varying appearance or shape of object. In this paper, we explore the applications of kernel based online learning methods in dealing with the above problems. We evaluate the performance of recently developed kernel based online algorithms combined with the state-of-the-art local shape feature descriptor. We perform the experimental evaluation on our dataset. The experimental results demonstrate that the online algorithms can be highly accurate to the problem of moving object classification.
{"title":"A comparison study on kernel based online learning for moving object classification","authors":"Xin Zhao, Kaiqi Huang, T. Tan","doi":"10.1109/IVSURV.2011.6157014","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157014","url":null,"abstract":"Most visual surveillance and video understanding systems require knowledge of categories of objects in the scene. One of the key challenges is to be able to classify any object in a real-time procedure in spite of changes in the scene over time and the varying appearance or shape of object. In this paper, we explore the applications of kernel based online learning methods in dealing with the above problems. We evaluate the performance of recently developed kernel based online algorithms combined with the state-of-the-art local shape feature descriptor. We perform the experimental evaluation on our dataset. The experimental results demonstrate that the online algorithms can be highly accurate to the problem of moving object classification.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"94 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":"127109947","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.2011.6157035
Huafeng Wang, Yunhong Wang, Zhaoxiang Zhang, Fan Wang, Jin Huang
A novel framework for unsupervised multi-faces tracking and recognition is built on Detection-Tracking-Recognition (DTR) approach. This framework proposed a hybrid face detector for real-time face tracking which is robust to occlusions and posture changes. Faces acquired during unsupervised detection stage will be further processed by SIFT operator in order to cluster face sequence into certain groups. After that, the relevant faces are put together which is of much importance for face recognition in videos. The framework is validated on several videos collected in unconstrained condition (20min each.).The framework can track the face and automatically group a serial faces for a single human-being object in an unlabeled video robustly.
{"title":"A multi-faces tracking and recognition framework for surveillance system","authors":"Huafeng Wang, Yunhong Wang, Zhaoxiang Zhang, Fan Wang, Jin Huang","doi":"10.1109/IVSURV.2011.6157035","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157035","url":null,"abstract":"A novel framework for unsupervised multi-faces tracking and recognition is built on Detection-Tracking-Recognition (DTR) approach. This framework proposed a hybrid face detector for real-time face tracking which is robust to occlusions and posture changes. Faces acquired during unsupervised detection stage will be further processed by SIFT operator in order to cluster face sequence into certain groups. After that, the relevant faces are put together which is of much importance for face recognition in videos. The framework is validated on several videos collected in unconstrained condition (20min each.).The framework can track the face and automatically group a serial faces for a single human-being object in an unlabeled video robustly.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"82 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":"115203115","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.2011.6157018
Qi Dong, Guangping Xu, Liu Jing, Yanbing Xue, Hua Zhang
The contour analysis and identification are the important aspects in visual surveillance research. The paper proposes a fuzzy identification method of contours. First, according to the description of a contour based on the chain-code method, the proposed method utilizes the statistical features of contours including the chain-code entropy and chain-code space distribution entropies, from which the feature vector of a contour is composed. Then, the method generates the contour pattern from some contour samples and uses the approaching principle to identify a contour. Since our method integrates effectively multiple statistical feathers of its chain-code and employs a fuzzy pattern recognition technique, the experiments show quantitatively that it can achieve good results from various metrics.
{"title":"A fuzzy identification method of contours based on chain-code features","authors":"Qi Dong, Guangping Xu, Liu Jing, Yanbing Xue, Hua Zhang","doi":"10.1109/IVSURV.2011.6157018","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157018","url":null,"abstract":"The contour analysis and identification are the important aspects in visual surveillance research. The paper proposes a fuzzy identification method of contours. First, according to the description of a contour based on the chain-code method, the proposed method utilizes the statistical features of contours including the chain-code entropy and chain-code space distribution entropies, from which the feature vector of a contour is composed. Then, the method generates the contour pattern from some contour samples and uses the approaching principle to identify a contour. Since our method integrates effectively multiple statistical feathers of its chain-code and employs a fuzzy pattern recognition technique, the experiments show quantitatively that it can achieve good results from various metrics.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"37 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":"117344004","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.2011.6157016
Guowen Ma, Jinfeng Yang
Shadows in images always cause problems to computer visual tasks, so how to remove shadow is an important topic of image processing. In this paper, we propose a new shadow removal method based on Retinex theory. We firstly use the gradient edge detection combined with 1-D illumination invariant image to detect the shadow area, then remove the shadow with Retinex algorithm, finally adjust the brightness of the shadow area. By transforming the RGB image into HSV space, we compute the average brightness of the non-shadow area in original image and shadow-free image, adjust the brightness of the image to coordinate with the original image. The experiment results show that our method can provide a good effect in shadow images.
{"title":"Shadow removal using Retinex theory","authors":"Guowen Ma, Jinfeng Yang","doi":"10.1109/IVSURV.2011.6157016","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157016","url":null,"abstract":"Shadows in images always cause problems to computer visual tasks, so how to remove shadow is an important topic of image processing. In this paper, we propose a new shadow removal method based on Retinex theory. We firstly use the gradient edge detection combined with 1-D illumination invariant image to detect the shadow area, then remove the shadow with Retinex algorithm, finally adjust the brightness of the shadow area. By transforming the RGB image into HSV space, we compute the average brightness of the non-shadow area in original image and shadow-free image, adjust the brightness of the image to coordinate with the original image. The experiment results show that our method can provide a good effect in shadow images.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"1 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":"121486942","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.2011.6157034
Shengyin Wu, Shiqi Yu, Wensheng Chen
We investigate pedestrian detection in depth images. Unlike pedestrian detection in intensity images, pedestrian detection in depth images can reduce the effect of complex background and illumination variation. We propose a new feature descriptor, Histogram of Depth Difference(HDD), for this task. The proposed HDD feature descriptor can describe the depth variance in a local region as Histogram of Oriented Gradients(HOG) describes local texture cues. To evaluate pedestrian detection in depth images, we also collected a large dataset, which contains not only depth images but also the synchronized intensity images. There are 4673 pedestrian samples in it. Our experimental results show that detecting pedestrians in depth images is feasible. We also fuse the HDD feature from depth images and HOG from intensity images. The fused feature gives an encouraging detection rate of 99.12% at FPPW=10−4.
我们研究了深度图像中的行人检测。与强度图像中的行人检测不同，深度图像中的行人检测可以减少复杂背景和光照变化的影响。为此我们提出了一种新的特征描述符——深度差直方图(Histogram of Depth Difference, HDD)。提出的HDD特征描述符可以用直方图定向梯度(HOG)描述局部纹理线索来描述局部区域的深度方差。为了评估深度图像中的行人检测，我们还收集了一个大型数据集，该数据集不仅包含深度图像，还包含同步强度图像。其中有4673个行人样本。实验结果表明，在深度图像中检测行人是可行的。我们还融合了深度图像的HDD特征和强度图像的HOG特征。在FPPW=10−4时，融合特征的检出率达到了令人鼓舞的99.12%。
{"title":"An attempt to pedestrian detection in depth images","authors":"Shengyin Wu, Shiqi Yu, Wensheng Chen","doi":"10.1109/IVSURV.2011.6157034","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157034","url":null,"abstract":"We investigate pedestrian detection in depth images. Unlike pedestrian detection in intensity images, pedestrian detection in depth images can reduce the effect of complex background and illumination variation. We propose a new feature descriptor, Histogram of Depth Difference(HDD), for this task. The proposed HDD feature descriptor can describe the depth variance in a local region as Histogram of Oriented Gradients(HOG) describes local texture cues. To evaluate pedestrian detection in depth images, we also collected a large dataset, which contains not only depth images but also the synchronized intensity images. There are 4673 pedestrian samples in it. Our experimental results show that detecting pedestrians in depth images is feasible. We also fuse the HDD feature from depth images and HOG from intensity images. The fused feature gives an encouraging detection rate of 99.12% at FPPW=10−4.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"3 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":"128944437","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.2011.6157020
Xiang Xiang
Long-term robust visual tracking is still a challenge, primarily due to the appearance changes of the scene and target. In this paper, we briefly review the recent progress in image representation, appearance model and motion model for building a general tracking system. The models reviewed here are basic enough to be applicable for tracking either single target or multiple targets. Special attention has been paid to the on-line adaptation of appearance model, a hot topic in the recent. Its key techniques have been discussed, such as classifier issue, on-line manner, sample selection and drifting problem. We notice that the recent state-of-the-art performances are generally given by a class of on-line boosting methods or ‘tracking-by-detection’ methods (e.g. OnlineBoost, SemiBoost, MIL-Track, TLD, etc.). Therefore, we validate them together with typical traditional methods (e.g. template matching, Mean Shift, optical flow, particle filter, FragTrack) on a challenging sequence for single person tracking. Qualitative comparison results are presented.
{"title":"A brief review on visual tracking methods","authors":"Xiang Xiang","doi":"10.1109/IVSURV.2011.6157020","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157020","url":null,"abstract":"Long-term robust visual tracking is still a challenge, primarily due to the appearance changes of the scene and target. In this paper, we briefly review the recent progress in image representation, appearance model and motion model for building a general tracking system. The models reviewed here are basic enough to be applicable for tracking either single target or multiple targets. Special attention has been paid to the on-line adaptation of appearance model, a hot topic in the recent. Its key techniques have been discussed, such as classifier issue, on-line manner, sample selection and drifting problem. We notice that the recent state-of-the-art performances are generally given by a class of on-line boosting methods or ‘tracking-by-detection’ methods (e.g. OnlineBoost, SemiBoost, MIL-Track, TLD, etc.). Therefore, we validate them together with typical traditional methods (e.g. template matching, Mean Shift, optical flow, particle filter, FragTrack) on a challenging sequence for single person tracking. Qualitative comparison results are presented.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"23 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":"134164410","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}
According to the technique requirements of a video monitoring and state recognition system for complex equipments, practical video image analysis and recognition algorithms are designed in this paper. Algorithms including dynamic registration of panel and precise adjustment of component based on Normalized Product Correlation (NPC), operation change detection based on Sobel edge strength dynamic analysis, multi-state component recognition based on Laplacian edge strength NPC matching, nixie tube reading based on NMI feature, meanwhile, the basic steps of these algorithms are given. The application experiments demonstrate the efficiency of the proposed algorithms.
{"title":"Design of algorithms for video monitoring and state recognition system for complex equipments","authors":"Xiao-gang Yang, Chuan Li, Bin-wen Chen, Fei Meng, Zhaohui Xia","doi":"10.1109/IVSURV.2011.6157017","DOIUrl":"https://doi.org/10.1109/IVSURV.2011.6157017","url":null,"abstract":"According to the technique requirements of a video monitoring and state recognition system for complex equipments, practical video image analysis and recognition algorithms are designed in this paper. Algorithms including dynamic registration of panel and precise adjustment of component based on Normalized Product Correlation (NPC), operation change detection based on Sobel edge strength dynamic analysis, multi-state component recognition based on Laplacian edge strength NPC matching, nixie tube reading based on NMI feature, meanwhile, the basic steps of these algorithms are given. The application experiments demonstrate the efficiency of the proposed algorithms.","PeriodicalId":141829,"journal":{"name":"2011 Third Chinese Conference on Intelligent Visual Surveillance","volume":"114 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":"128138526","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：