Pub Date : 2012-04-18DOI: 10.1109/SIU.2012.6204665
Sedat Telçeken, M. Dogan, Ö. N. Gerek
This study aims to improve the efficiency of the hybrid (lifting and motion compensation) video compression method by incorporating variable block size for block matching. That hybrid method first separates even and odd numbered video frames. Then the even frames are predicted using a temporal edge adapted idea, which was previously proposed for 2D image coding. To further improve the efficiency, a novel symmetric block matching pre-step is applied. This method was observed to perform well in test video sequences. In this work, to further improve the block matching representation efficiency, a variable size block idea is proposed. The new proposed method compares neighboring blocks in terms of motion vectors, and merges them to a larger size if the directions and sizes of motions are common. The proposed method yields PSNR values greater than that of MPEG2 and closer to that of H.264 in the fixed compression ratios.
{"title":"Motion compensated video coder based on block matching method with variable-sized blocks","authors":"Sedat Telçeken, M. Dogan, Ö. N. Gerek","doi":"10.1109/SIU.2012.6204665","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204665","url":null,"abstract":"This study aims to improve the efficiency of the hybrid (lifting and motion compensation) video compression method by incorporating variable block size for block matching. That hybrid method first separates even and odd numbered video frames. Then the even frames are predicted using a temporal edge adapted idea, which was previously proposed for 2D image coding. To further improve the efficiency, a novel symmetric block matching pre-step is applied. This method was observed to perform well in test video sequences. In this work, to further improve the block matching representation efficiency, a variable size block idea is proposed. The new proposed method compares neighboring blocks in terms of motion vectors, and merges them to a larger size if the directions and sizes of motions are common. The proposed method yields PSNR values greater than that of MPEG2 and closer to that of H.264 in the fixed compression ratios.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114351733","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204596
Serdar Çakır, T. Aytaç, A. Yildirim, S. Beheshti, O. Gerek, A. Çetin
Features extracted at salient points in the image are used to construct region covariance descriptor (RCD) for target tracking purposes. In the classical approach, the RCD is computed by using the features at each pixel location and thus, increases the computational cost in the scenarios where large targets are tracked. The approach in which the features at each pixel location are used, is redundant in cases where image statistics do not change significantly between neighboring pixels. Furthermore, this may decrease the tracking accuracy while tracking large targets which have background dominating structures. In the proposed approach, the salient points are extracted via the Shi and Tomasi's minimum eigenvalue method and a descriptor based target tracking structure is constructed based on the features extracted only at these salient points. Experimental results indicate that the proposed method provides comparable and in some cases even better tracking results compared to the classical method while providing a computationally more efficient structure.
{"title":"Region covariance descriptors calculated over the salient points for target tracking","authors":"Serdar Çakır, T. Aytaç, A. Yildirim, S. Beheshti, O. Gerek, A. Çetin","doi":"10.1109/SIU.2012.6204596","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204596","url":null,"abstract":"Features extracted at salient points in the image are used to construct region covariance descriptor (RCD) for target tracking purposes. In the classical approach, the RCD is computed by using the features at each pixel location and thus, increases the computational cost in the scenarios where large targets are tracked. The approach in which the features at each pixel location are used, is redundant in cases where image statistics do not change significantly between neighboring pixels. Furthermore, this may decrease the tracking accuracy while tracking large targets which have background dominating structures. In the proposed approach, the salient points are extracted via the Shi and Tomasi's minimum eigenvalue method and a descriptor based target tracking structure is constructed based on the features extracted only at these salient points. Experimental results indicate that the proposed method provides comparable and in some cases even better tracking results compared to the classical method while providing a computationally more efficient structure.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114572985","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204706
M. C. Sahingil, Yakup S. Özkazanç
There are lots of regions in human DNA (deoxy-ribo-nucleic acid) sequences which contain repetitive patterns. In this paper, the visualization of repetitive regions of DNA sequences via Short Time Fourier Transform is investigated.
{"title":"Visualization of repetitive DNA sequence regions via Short Time Fourier Transform","authors":"M. C. Sahingil, Yakup S. Özkazanç","doi":"10.1109/SIU.2012.6204706","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204706","url":null,"abstract":"There are lots of regions in human DNA (deoxy-ribo-nucleic acid) sequences which contain repetitive patterns. In this paper, the visualization of repetitive regions of DNA sequences via Short Time Fourier Transform is investigated.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114704995","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204805
Perihan Isinsu Akçetin, S. Ergen, T. M. Sezgin
In this paper, a system of inertial sensors is presented for observing and analyzing the rowing technique of the athletes. Monitoring of the oarsman by coaches is a key element for improving the performance and technique, and health of the athlete. However, reliability of this type of observation is limited by human ability and availability, hence lacks a scientific standardization. Therefore, in order to improve the accuracy and quality of the feedback given by the coaches on the technique of the sportsman and maybe even to eliminate the necessity for a human trainer for good, a system supported by inertial sensors is offered. The prototype presented by this work promises to monitor and analyze the kinematics of major body parts during rowing. Data are collected from the inertial sensors placed on lower back, femur and forearm when either professional or amateur rowers are using ergometer. Due to its features of small size, lightweight, bluetooth connection, lowpower usage and integrated 3 axis accelerometer & 3 axis gyroscope, SHIMMER sensor nodes has been used during these sessions. After data have been extracted, they were processed and an HMM has been created using the correct rowing technique data. Then, the worse rowing cases were compared using the built HMM.
{"title":"HMM based inertial sensor system for coaching of rowing activity","authors":"Perihan Isinsu Akçetin, S. Ergen, T. M. Sezgin","doi":"10.1109/SIU.2012.6204805","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204805","url":null,"abstract":"In this paper, a system of inertial sensors is presented for observing and analyzing the rowing technique of the athletes. Monitoring of the oarsman by coaches is a key element for improving the performance and technique, and health of the athlete. However, reliability of this type of observation is limited by human ability and availability, hence lacks a scientific standardization. Therefore, in order to improve the accuracy and quality of the feedback given by the coaches on the technique of the sportsman and maybe even to eliminate the necessity for a human trainer for good, a system supported by inertial sensors is offered. The prototype presented by this work promises to monitor and analyze the kinematics of major body parts during rowing. Data are collected from the inertial sensors placed on lower back, femur and forearm when either professional or amateur rowers are using ergometer. Due to its features of small size, lightweight, bluetooth connection, lowpower usage and integrated 3 axis accelerometer & 3 axis gyroscope, SHIMMER sensor nodes has been used during these sessions. After data have been extracted, they were processed and an HMM has been created using the correct rowing technique data. Then, the worse rowing cases were compared using the built HMM.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115073050","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204693
O. Yıldız, M. Tez, H. Ş. Bilge, M. Ali Akcayol, I. Güler
Breast cancer can be fatal and so it is very dangerous. Early diagnosis of breast cancer has been playing very important role on treatment of the disease. Recently, gene technology has been widely used in cancer diagnosis. A microarray is a tool for analyzing gene expression. Microarray data usually contain thousands of genes and a small number of samples. Although, most of them are irrelevant or insignificant to a clinical diagnosis. It is very difficult to obtain a satisfactory classification result by machine learning techniques because of both the curse-of dimensionality problem and the overfitting problem. Therefore, feature selection plays a crucial role in microarray analysis. In this paper, significant biomarker genes for diagnosis have been identified by feature selection. We attempted to use these markers for the classification of breast cancer. Subsequently, SVM was also used to verify the classification rate of genes selected by feature selection. The classification rate of SVM reaches to 82.69% when using selected genes.
{"title":"Gene selection for breast cancer","authors":"O. Yıldız, M. Tez, H. Ş. Bilge, M. Ali Akcayol, I. Güler","doi":"10.1109/SIU.2012.6204693","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204693","url":null,"abstract":"Breast cancer can be fatal and so it is very dangerous. Early diagnosis of breast cancer has been playing very important role on treatment of the disease. Recently, gene technology has been widely used in cancer diagnosis. A microarray is a tool for analyzing gene expression. Microarray data usually contain thousands of genes and a small number of samples. Although, most of them are irrelevant or insignificant to a clinical diagnosis. It is very difficult to obtain a satisfactory classification result by machine learning techniques because of both the curse-of dimensionality problem and the overfitting problem. Therefore, feature selection plays a crucial role in microarray analysis. In this paper, significant biomarker genes for diagnosis have been identified by feature selection. We attempted to use these markers for the classification of breast cancer. Subsequently, SVM was also used to verify the classification rate of genes selected by feature selection. The classification rate of SVM reaches to 82.69% when using selected genes.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116417694","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204676
B. Yildirim, A. Bayri, Gokhan Gok
Finding the direction of the emitters is a problem to be solved in many areas. One of the methods used for that purpose is that using a rotating directional antenna and obtaining the direction estimate by processing the measured signal amplitudes. Direction estimation, in principle, is based on the own antenna pattern modulation on the measured signal. In the paper, direction estimates of one or more emitters are found with two different methods (matched filtering and Wiener deconvolution), and performances (signal-to-noise ratio, changing emitter power, etc.) of these methods are studied.
{"title":"Direction finding with a rotating antenna","authors":"B. Yildirim, A. Bayri, Gokhan Gok","doi":"10.1109/SIU.2012.6204676","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204676","url":null,"abstract":"Finding the direction of the emitters is a problem to be solved in many areas. One of the methods used for that purpose is that using a rotating directional antenna and obtaining the direction estimate by processing the measured signal amplitudes. Direction estimation, in principle, is based on the own antenna pattern modulation on the measured signal. In the paper, direction estimates of one or more emitters are found with two different methods (matched filtering and Wiener deconvolution), and performances (signal-to-noise ratio, changing emitter power, etc.) of these methods are studied.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123930197","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204583
Salim Sirtkaya, Aydin Alatan
This paper proposes an efficient plane hypothesis matching technique for 3D mapping of urban environments using images obtained from a moving monocular camera. The algorithm is based on the assumption that urban environments are generally composed of buildings that have planar facades, and these facades are placed in the direction of gravity. A sparse 3D point cloud of the imaged scene is obtained using the classical Structure from Motion technique, and then the plane hypotheses are obtained by running an iterative Hough Transform on the 2D point set that is obtained from the projection of these 3D points in the direction of gravity. Superpixels are preferred instead of pixels for matching the image to the plane hypotheses. The superpixels are assigned to the plane hypotheses using their 3D point associations. As a result, a dense depth map of the urban scene is constructed successfully by means of the planar patches.
本文提出了一种利用运动单目相机图像进行城市环境三维映射的有效平面假设匹配技术。该算法基于这样的假设:城市环境通常由具有平面立面的建筑物组成，这些立面被放置在重力方向上。利用经典的Structure from Motion技术获得图像场景的稀疏三维点云，然后对这些三维点在重力方向上的投影得到的二维点集进行迭代霍夫变换，得到平面假设。在将图像与平面假设匹配时，首选超像素而不是像素。超像素使用它们的3D点关联分配给平面假设。结果表明，利用平面斑块成功构建了密集的城市场景深度图。
{"title":"3D modeling of urban areas using plane hypotheses","authors":"Salim Sirtkaya, Aydin Alatan","doi":"10.1109/SIU.2012.6204583","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204583","url":null,"abstract":"This paper proposes an efficient plane hypothesis matching technique for 3D mapping of urban environments using images obtained from a moving monocular camera. The algorithm is based on the assumption that urban environments are generally composed of buildings that have planar facades, and these facades are placed in the direction of gravity. A sparse 3D point cloud of the imaged scene is obtained using the classical Structure from Motion technique, and then the plane hypotheses are obtained by running an iterative Hough Transform on the 2D point set that is obtained from the projection of these 3D points in the direction of gravity. Superpixels are preferred instead of pixels for matching the image to the plane hypotheses. The superpixels are assigned to the plane hypotheses using their 3D point associations. As a result, a dense depth map of the urban scene is constructed successfully by means of the planar patches.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"42 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125900276","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204699
Umut Simsekli, Y. K. Yilmaz, A. Cemgil
Generalized Coupled Tensor Factorization (GCTF) is a recently proposed algorithmic framework for simultaneously estimating tensor factorization models where several tensors can share a set of latent factors. This paper presents two models in this framework for transcribing polyphonic piano pieces. The first model is based on Non-negative Matrix Factorization where the coupling provides the spectral information to the model. As an extension to the first model, the second model incorporates temporal and harmonic information by taking a rough, incomplete transciption of the piece as input. Incorporating harmonic knowledge improves the transcription quality as the the experimental results show that we get around 23 % F-measure improvement on real piano data.
{"title":"Coupled tensor factorization models for polyphonic music transcription","authors":"Umut Simsekli, Y. K. Yilmaz, A. Cemgil","doi":"10.1109/SIU.2012.6204699","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204699","url":null,"abstract":"Generalized Coupled Tensor Factorization (GCTF) is a recently proposed algorithmic framework for simultaneously estimating tensor factorization models where several tensors can share a set of latent factors. This paper presents two models in this framework for transcribing polyphonic piano pieces. The first model is based on Non-negative Matrix Factorization where the coupling provides the spectral information to the model. As an extension to the first model, the second model incorporates temporal and harmonic information by taking a rough, incomplete transciption of the piece as input. Incorporating harmonic knowledge improves the transcription quality as the the experimental results show that we get around 23 % F-measure improvement on real piano data.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"96 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126005634","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204575
Burcu Kir Savaş, Cemil Öz, Ali Gülbağ
Plants play a crucial role in terms of the lives of human and other creatures since the existence of the universe. Despite the studies of plant scientists, there are many undiscovered and unidentified species in our environment. This paper is aimed to add the leaves, whose images have been clearly attained, to the system and to provide a proper analysis of those leaves. The images could be either the ones taken before or the ones obtained by means of a camera that is connected transiently. Leaf images went through pretreatment phases first, and then their features were extracted. Finally, classification processing was accomplished by using K-NN algorithm. The System is working successfully.
{"title":"Leaf recognition using K-NN classification algorithm","authors":"Burcu Kir Savaş, Cemil Öz, Ali Gülbağ","doi":"10.1109/SIU.2012.6204575","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204575","url":null,"abstract":"Plants play a crucial role in terms of the lives of human and other creatures since the existence of the universe. Despite the studies of plant scientists, there are many undiscovered and unidentified species in our environment. This paper is aimed to add the leaves, whose images have been clearly attained, to the system and to provide a proper analysis of those leaves. The images could be either the ones taken before or the ones obtained by means of a camera that is connected transiently. Leaf images went through pretreatment phases first, and then their features were extracted. Finally, classification processing was accomplished by using K-NN algorithm. The System is working successfully.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124601480","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 : 2012-04-18DOI: 10.1109/SIU.2012.6204833
U. Konur, F. Gürgen, F. Varol
In this study, we use ultrasound (US) imaging modality frequently employed in prenatal diagnosis and axial skull images used primarily in the examination of fetal neural tubes and work on the segmentation of skull (and brain) structures. The segmentation performance of the mentioned structures is vital in that, applications such as automatic diagnosis systems can provide better feature extraction and classification performance with the aid of such a preprocessing. Our approach works with the principles of coarsely localizing the skull and brain structures present in US images acquired in transverse sections of fetal skulls using model (ellipse) fitting and successively obtaining more accurate segmentation with Active Appearance Models, which is a learning-based segmentation algorithm.
{"title":"Segmentation of fetal skulls using ellipse fitting and active appearance models","authors":"U. Konur, F. Gürgen, F. Varol","doi":"10.1109/SIU.2012.6204833","DOIUrl":"https://doi.org/10.1109/SIU.2012.6204833","url":null,"abstract":"In this study, we use ultrasound (US) imaging modality frequently employed in prenatal diagnosis and axial skull images used primarily in the examination of fetal neural tubes and work on the segmentation of skull (and brain) structures. The segmentation performance of the mentioned structures is vital in that, applications such as automatic diagnosis systems can provide better feature extraction and classification performance with the aid of such a preprocessing. Our approach works with the principles of coarsely localizing the skull and brain structures present in US images acquired in transverse sections of fetal skulls using model (ellipse) fitting and successively obtaining more accurate segmentation with Active Appearance Models, which is a learning-based segmentation algorithm.","PeriodicalId":256154,"journal":{"name":"2012 20th Signal Processing and Communications Applications Conference (SIU)","volume":"124 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124650023","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：