ABSTRACT In this study, a superresolution method using registered, noisy and down sampled images ... more ABSTRACT In this study, a superresolution method using registered, noisy and down sampled images is presented. Maximum a Posteriori (MAP) method, one of the statistical pixel domain approaches, is used as the superresolution algorithm. Performances of different data fidelity terms and regularization terms used in the literature are shown. In most of the applications the effects that degrade the image frame are assumed to be known completely or known limited. In this application, the performances of the several methods used to find the amount of blur caused by the unfocussed camera lenses are shown and the best method results are used in the superrresolution algorithm. In this way, the error value of superresolved image is decreased.
Automatic detection of targets from far distances is a very challenging problem. Background clutt... more Automatic detection of targets from far distances is a very challenging problem. Background clutter and small target size are the main difficulties which should be solved while reaching a high detection performance as well as a low computational load. The pre-processing, detection and post-processing approaches are very effective on the final results. In this study, first of all, various methods in the literature were evaluated separately for each of these stages using the simulated test scenarios. Then, a full system of detection was constructed among available solutions which resulted in the best performance in terms of detection. However, although a precision rate as 100% was reached, the recall values stayed low around 25-45%. Finally, a post-processing method was proposed which increased the recall value while keeping the precision at 100%. The proposed post-processing method, which is based on local operations, increased the recall value to 65-95% in all test scenarios.
The main aim of this study was to compare and discuss the effective brain connectivity maps of dy... more The main aim of this study was to compare and discuss the effective brain connectivity maps of dyslexic and control groups in terms of differences and similarities. The differences and sismilarities that may be found in the study, could provide information for the related future studies. There were total of 58 subjects which were 27 control and 31 dyslexics data in this study. All the data according to the groups were averaged with respect to time and group data were generated. Then, effective connectivity maps for both dyslexic and control groups were constructed using the Dynamic Bayesian Networks algorithms. The symmetric connectivities which we call strong connections between electrode pairs were extracted and mainly used for discussion. It was observed that there are differences in effective connectivity maps of both groups in terms of connections' density and directions. The results should also be checked by brain anatomy experts anatomically and by comparing with literature of medicine, their validity and usabilities could be tested.
Pre-processing, thresholding and post-processing stages are very important especially for very sm... more Pre-processing, thresholding and post-processing stages are very important especially for very small target detection from infrared images. The effects of these stages to the final detection performance are measured in this study. Various methods for each stage are compared based on the final detection performance, which is defined by precision and recall values. Among various methods, the best method for each stage is selected and proved. For the pre-processing stage, local block based methods perform the best, nearly for all thresholding methods. The best thresholding method is chosen as the one, which does not need any user defined parameter. Finally, the post processing method, which is suitable for the best performing pre-processing and thesholding methods is selected.
International Conference on Pattern Recognition, Aug 23, 2004
In sparse stereo [6], [12], [13], [14], [18], distinctive image features are extracted and corres... more In sparse stereo [6], [12], [13], [14], [18], distinctive image features are extracted and corresponding pairs are matched using a feature-based criteria. To be successful for stereo applications, local features must be robust to image deformations such as noise, rotation, scale and brightness changes. Jenkin, Jepson and Fleet [7], [8], [9] and Sanger [19] describe promising methods based on the output phase behavior of band-pass Gabor filters. Recently, Carneiro and Jepson show that the phase information provided by steerable filters is often locally stable with respect ...
Retinal vessel segmentation and extracting features such as tortuosity, width, length related to ... more Retinal vessel segmentation and extracting features such as tortuosity, width, length related to those vessels can be used in diagnosis, treatment and screening of many diseases such as retinopathy of prematurity, hypertension and diabetes. Therefore, automatic segmentation of vessels by computers will make the analysis of those diseases easier and will help during the screening, diagnosis and treatment processes. In this study, a solution based on convolutional neural networks (CNN) is proposed for automatic segmentation of retinal vessels. The proposed CNN model is tested on DRIVE dataset and a better performance than literature is achieved.
Lecture notes in electrical engineering, Aug 18, 2010
Gender and Age Groups Classifications for Semantic Annotation of Videos Gökhan Yaprakkaya 1, Niha... more Gender and Age Groups Classifications for Semantic Annotation of Videos Gökhan Yaprakkaya 1, Nihan Kesim Cicekli 1, İlkay Ulusoy 2 1 Department of ... were satisfactory (see Tables 1 and 2). The running time of our method changes between 2 and 4 ms on Intel (R) Core (TM ...
ABSTRACT Full Search video stabilization method is implemented on FPGA to realize its real time p... more ABSTRACT Full Search video stabilization method is implemented on FPGA to realize its real time performance. Also, the method is implemented and tested in MATLAB. FPGA results are compared with MATLAB's to see the accuracy performance. The input video is PAL which frame period is 40 ms. The FPGA implementation is capable of producing new stabilization data at every PAL frame which allows the implementation to be classified as real time. Simulation and hardware tests show that FPGA implementation can reach the MATLAB accuracy performance.
ABSTRACT In this study, a superresolution method using registered, noisy and down sampled images ... more ABSTRACT In this study, a superresolution method using registered, noisy and down sampled images is presented. Maximum a Posteriori (MAP) method, one of the statistical pixel domain approaches, is used as the superresolution algorithm. Performances of different data fidelity terms and regularization terms used in the literature are shown. In most of the applications the effects that degrade the image frame are assumed to be known completely or known limited. In this application, the performances of the several methods used to find the amount of blur caused by the unfocussed camera lenses are shown and the best method results are used in the superrresolution algorithm. In this way, the error value of superresolved image is decreased.
Automatic detection of targets from far distances is a very challenging problem. Background clutt... more Automatic detection of targets from far distances is a very challenging problem. Background clutter and small target size are the main difficulties which should be solved while reaching a high detection performance as well as a low computational load. The pre-processing, detection and post-processing approaches are very effective on the final results. In this study, first of all, various methods in the literature were evaluated separately for each of these stages using the simulated test scenarios. Then, a full system of detection was constructed among available solutions which resulted in the best performance in terms of detection. However, although a precision rate as 100% was reached, the recall values stayed low around 25-45%. Finally, a post-processing method was proposed which increased the recall value while keeping the precision at 100%. The proposed post-processing method, which is based on local operations, increased the recall value to 65-95% in all test scenarios.
The main aim of this study was to compare and discuss the effective brain connectivity maps of dy... more The main aim of this study was to compare and discuss the effective brain connectivity maps of dyslexic and control groups in terms of differences and similarities. The differences and sismilarities that may be found in the study, could provide information for the related future studies. There were total of 58 subjects which were 27 control and 31 dyslexics data in this study. All the data according to the groups were averaged with respect to time and group data were generated. Then, effective connectivity maps for both dyslexic and control groups were constructed using the Dynamic Bayesian Networks algorithms. The symmetric connectivities which we call strong connections between electrode pairs were extracted and mainly used for discussion. It was observed that there are differences in effective connectivity maps of both groups in terms of connections' density and directions. The results should also be checked by brain anatomy experts anatomically and by comparing with literature of medicine, their validity and usabilities could be tested.
Pre-processing, thresholding and post-processing stages are very important especially for very sm... more Pre-processing, thresholding and post-processing stages are very important especially for very small target detection from infrared images. The effects of these stages to the final detection performance are measured in this study. Various methods for each stage are compared based on the final detection performance, which is defined by precision and recall values. Among various methods, the best method for each stage is selected and proved. For the pre-processing stage, local block based methods perform the best, nearly for all thresholding methods. The best thresholding method is chosen as the one, which does not need any user defined parameter. Finally, the post processing method, which is suitable for the best performing pre-processing and thesholding methods is selected.
International Conference on Pattern Recognition, Aug 23, 2004
In sparse stereo [6], [12], [13], [14], [18], distinctive image features are extracted and corres... more In sparse stereo [6], [12], [13], [14], [18], distinctive image features are extracted and corresponding pairs are matched using a feature-based criteria. To be successful for stereo applications, local features must be robust to image deformations such as noise, rotation, scale and brightness changes. Jenkin, Jepson and Fleet [7], [8], [9] and Sanger [19] describe promising methods based on the output phase behavior of band-pass Gabor filters. Recently, Carneiro and Jepson show that the phase information provided by steerable filters is often locally stable with respect ...
Retinal vessel segmentation and extracting features such as tortuosity, width, length related to ... more Retinal vessel segmentation and extracting features such as tortuosity, width, length related to those vessels can be used in diagnosis, treatment and screening of many diseases such as retinopathy of prematurity, hypertension and diabetes. Therefore, automatic segmentation of vessels by computers will make the analysis of those diseases easier and will help during the screening, diagnosis and treatment processes. In this study, a solution based on convolutional neural networks (CNN) is proposed for automatic segmentation of retinal vessels. The proposed CNN model is tested on DRIVE dataset and a better performance than literature is achieved.
Lecture notes in electrical engineering, Aug 18, 2010
Gender and Age Groups Classifications for Semantic Annotation of Videos Gökhan Yaprakkaya 1, Niha... more Gender and Age Groups Classifications for Semantic Annotation of Videos Gökhan Yaprakkaya 1, Nihan Kesim Cicekli 1, İlkay Ulusoy 2 1 Department of ... were satisfactory (see Tables 1 and 2). The running time of our method changes between 2 and 4 ms on Intel (R) Core (TM ...
ABSTRACT Full Search video stabilization method is implemented on FPGA to realize its real time p... more ABSTRACT Full Search video stabilization method is implemented on FPGA to realize its real time performance. Also, the method is implemented and tested in MATLAB. FPGA results are compared with MATLAB's to see the accuracy performance. The input video is PAL which frame period is 40 ms. The FPGA implementation is capable of producing new stabilization data at every PAL frame which allows the implementation to be classified as real time. Simulation and hardware tests show that FPGA implementation can reach the MATLAB accuracy performance.
Uploads