Estimation of surface roughness is an important quality measure for many applications including o... more Estimation of surface roughness is an important quality measure for many applications including optics, polymers, semiconductor etc. In this paper, we propose to estimate surface roughness using image focus. We use polymers as test objects. The proposed method is an inexpensive and fast method based on Shape From Focus (SFF). We show that the results from SFF are encouraging for comparison of polymers using surface roughness. .
Mean Shift (MS) tracking using histogram features alone may cause inaccuracy in target localizati... more Mean Shift (MS) tracking using histogram features alone may cause inaccuracy in target localization. The problem becomes worst due to presence of mingled background features in target model representation. To improve MS target localization problem, this paper propose a spatiospectral technique. The true background features are identified in target model representation using spectral and spatial weighting and then a transformation is applied to minimize their effect in target model representation for localization improvement. The target localization is further improved by adjusting the MS estimated target position through edge based centroid re positioning. The paper also propose method of target model update for background weighted histogram based algorithms followed by weighted transformation through online feature consistency data. The proposed method is designed for single object tracking in complex scenarios and tested for comparative results with existing state of the art techniques. Experimental results on numerous challenging video sequences verify the significance of proposed technique in terms of robustness to complex background, occlusions, appearance changes, and similar color object avoidance.
Image registration is a spatial alignment of two or more images and essential technology in image... more Image registration is a spatial alignment of two or more images and essential technology in image fusion, surveillance. Unlike the electro-optics (EO) sensor, infrared (IR) sensor absorbs the radiation energy. The relation between IR and EO image of pixel intensity is sometimes similar or often reverse. To overcome this complicated problem, we propose block-based processing incorporating corner detector and descriptor. This method also uses the Gaussian pyramids to register EO/IR image. The proposed method provides more accurate registration results.
In this paper, we propose a pulmonary nodule segmentation method by using a 3D deformable model. ... more In this paper, we propose a pulmonary nodule segmentation method by using a 3D deformable model. First of all, the initial model is generated. The shape of the initial model is sphere which consists of triangle mashes. The model has internal energy and external energy. These energies are optimized to segment the nodule. Accordingly, the model is deformed through point normal vectors. After this process, the deformed models are evaluated. These processes are iterated until the model is converged.
With rapidly growing market for three-dimensional (3D) movies, 3D TVs, and 3D gaming, we may now ... more With rapidly growing market for three-dimensional (3D) movies, 3D TVs, and 3D gaming, we may now be entering a time called the "Era of 3D". The general public has also started to become comfortable with stereoscopic vision. However, there are concerns about the effects on the human body from continuously watching 3D images, for examples, visually induced motion sickness (VIMS), visual fatigue, and asthenopia. Although their mechanisms are still not fully understood, there is a great need for more knowledge about the effects of those products on users and guidelines
Abstract Three-dimensional object reconstructions is an active research area in digital imaging. ... more Abstract Three-dimensional object reconstructions is an active research area in digital imaging. In shape from focus approach, erroneous focus measurements result in inaccuracy in the depth map reconstruction of 2D object. Conventionally, to enhance the image focus volume, focus values are aggregated within a window, which is a linear filtering approach. Owing to the inherent limitation of linear process, optimal results may not be obtained. In order to overcome this limitation, a non-linear filtering approach is proposed to enhance the image focus volume for accurate depth estimation. The noisy focus values are restored in two steps. First, noisy focus values are detected using min–max operators. In order to increase the dynamic range between the minimum and the maximum focus values within the window, an appropriate power law function is designed. In second step, only the noisy measurements are replaced with the estimated ones. A refined depth map is obtained from the updated focus volume. This process continues until the difference between the previous and the current depth maps becomes very small. The performance of the proposed non-linear filtering approach is obtained for various synthetic and real objects. The results highlight the depth map estimates of the proposed approach more accurate while preserving object edges. Comparative analysis demonstrates the effectiveness of the proposed approach.
Radiologists are interested in finding the stage of cancer, so the patient can be treated and cur... more Radiologists are interested in finding the stage of cancer, so the patient can be treated and cured accordingly. This is possible by finding the type of abnormality to measure the severity of cancer in mammograms. CAD could provide them the option of better opinion about the type of abnormality. In this paper, we have proposed a novel method which can classify cancerous mammogram into six classes. Features are extracted from preprocessed images and passed through different classifiers to identify malignant mammograms and the results of winning algorithm that is Support Vector Machine (SVM) in this case are considered for next processing. Mammograms declared as malignant by SVM are divided into six classes. Again, binary classifier (SVM) is used for multi-classification using one against all technique for classification. Output of all classifiers is combined by max, median and mean rule. It has been noted that results are very much satisfactory and accuracy of classification of abnormalities is more than 96% in case of max rule. MIAS [47] data set is used for experimentation purpose.
Discrete cosine transform (DCT)-and wavelet-based schemes have been highly successful for image c... more Discrete cosine transform (DCT)-and wavelet-based schemes have been highly successful for image compression in the lossy mode. Further improvements in one or more stages of these schemes are frequently reported in literature. The input images to these methods ...
Abstract The technique to estimate the three-dimensional (3D) geometry of an object from a sequen... more Abstract The technique to estimate the three-dimensional (3D) geometry of an object from a sequence of images obtained at different focus settings is called shape from focus (SFF). In SFF, the measure of focus — sharpness — is the crucial part for final 3D shape estimation. However, it is difficult to compute accurate and precise focus value because of the noise presence during the image acquisition by imaging system. Various noise filters can be employed to tackle this problem, but they also remove the sharpness information in addition to the noise. In this paper, we propose a method based on mean shift algorithm to remove noise introduced by the imaging process while minimising loss of edges. We test the algorithm in the presence of Gaussian noise and impulse noise. Experimental results show that the proposed algorithm based on the mean shift algorithm provides better results than the traditional focus measures in the presence of the above mentioned two types of noise.
A new method for the three-dimensional shape recovery from image focus is proposed. The method is... more A new method for the three-dimensional shape recovery from image focus is proposed. The method is based on approximation of the Focussed Image Surface (FIS) by a piecewise curved surface which tracks the realistic FIS in image space. The piecewise ...
... Muhammad Bilal Ahmad and Tae-Sun Choi, SeniorMember IEEE ... [4] Muhammad Bilal Ahmad and Tae... more ... Muhammad Bilal Ahmad and Tae-Sun Choi, SeniorMember IEEE ... [4] Muhammad Bilal Ahmad and Tae-Sun Choi, A heuristic approach for finding best focused shape, IEEE Transactions on Circuit System for Video Technologies (to be published in April 2005 issue). II - 972
Abstract-When a stego image undergoes an attack, the embedded watermark is also distorted. This p... more Abstract-When a stego image undergoes an attack, the embedded watermark is also distorted. This pertains to the ability of a watermark holding to the host image and undergoing the same attack as the host image does. However, this attribute of the ...
We present a novel approach to developing Machine Learning (ML) based decoding models for extract... more We present a novel approach to developing Machine Learning (ML) based decoding models for extracting a watermark in the presence of attacks. Statistical characterization of the components of various frequency bands is exploited to allow blind extraction of the ...
... 316-317, 2001. [4] Y. K. Seong, Y. H. Choi, and T. S. Choi, Design and Implementation of Har... more ... 316-317, 2001. [4] Y. K. Seong, Y. H. Choi, and T. S. Choi, Design and Implementation of Hard Disk Drive Embedded Digital Satellite Receiver, IEEE International Symposium on Consumer Electronics, pp.168-173, 2001. [5] U. Gargi, R. Kasturi, and SH Strayer, Performance ...
This paper suggests a simple scheme in which the search pattern is divided into a number of secto... more This paper suggests a simple scheme in which the search pattern is divided into a number of sectors based on the spatio-temporal correlation information. In the first step five neighboring blocks are searched for finding the predicted motion vector. The predictive motion vector thus obtained is chosen as the initial search center. This predictive search center is found to be closer to the global minimum and thus decreases the effects of the monotonic error surface assumption and its impact on the motion field estimates. Secondly the prediction information is used to obtain the direction of predicted motion vector. Based on the direction of predicted motion vector the search area is divided into four sectors. Final search pattern is adaptive and depends on the sector selected and significantly reduces the computational complexity. Experiments show the speed improvement of the proposed algorithm as compared to other fast search algorithms, in addition the image quality measured in terms of PSNR also shows good results.
ABSTRACT This paper presents a fast algorithm for H.264 fractional motion estimation (ME). In H.2... more ABSTRACT This paper presents a fast algorithm for H.264 fractional motion estimation (ME). In H.264 ME is the most time consuming component. The ME process consists of two stages: integer pixel and fractional pixel search. To reduce the complexity of fractional pixel ME we propose a quadrant based directional fractional pixel ME algorithm that is based on the unimodal property of the fractional pixel error surface. Better computation reduction has been achieved by using this strategy. Experimental results show that as compared to fast sub-pel ME proposed in H.264, the proposed method can speedup fractional ME, with a negligible degradation in video quality.
ABSTRACT This paper suggests a simple scheme based on spatio-temporal neighborhood information fo... more ABSTRACT This paper suggests a simple scheme based on spatio-temporal neighborhood information for obtaining better estimates of motion vectors. The estimated motion vector is chosen as initial search center. This predictive search center is found closer to the global minimum. Based on the prediction, the algorithm also chooses between center-biased or uniform approach for slow or fast moving sequences. For final fine search quadrant selection algorithm is chosen that speeds up the process. Experimental results presented in this paper demonstrate the efficiency of the proposed approach.
Previous studies show that the motion vector distribution within the search window shows a center... more Previous studies show that the motion vector distribution within the search window shows a center biased behavior. Based on this fact, an advanced center biased search algorithm for fast block motion estimation has been proposed. The algorithm drastically reduces the computational complexity by strict application of the unimodal error surface assumption. Improved error performance has been achieved by an efficient center biased search strategy that improves the chances of getting the correct motion vector. The half stop technique has been adopted to speed up the block matching process. Experimental results show that proposed algorithm has improved performance as compared to the three step search algorithm. A good predicted image quality is also achieved
Estimation of surface roughness is an important quality measure for many applications including o... more Estimation of surface roughness is an important quality measure for many applications including optics, polymers, semiconductor etc. In this paper, we propose to estimate surface roughness using image focus. We use polymers as test objects. The proposed method is an inexpensive and fast method based on Shape From Focus (SFF). We show that the results from SFF are encouraging for comparison of polymers using surface roughness. .
Mean Shift (MS) tracking using histogram features alone may cause inaccuracy in target localizati... more Mean Shift (MS) tracking using histogram features alone may cause inaccuracy in target localization. The problem becomes worst due to presence of mingled background features in target model representation. To improve MS target localization problem, this paper propose a spatiospectral technique. The true background features are identified in target model representation using spectral and spatial weighting and then a transformation is applied to minimize their effect in target model representation for localization improvement. The target localization is further improved by adjusting the MS estimated target position through edge based centroid re positioning. The paper also propose method of target model update for background weighted histogram based algorithms followed by weighted transformation through online feature consistency data. The proposed method is designed for single object tracking in complex scenarios and tested for comparative results with existing state of the art techniques. Experimental results on numerous challenging video sequences verify the significance of proposed technique in terms of robustness to complex background, occlusions, appearance changes, and similar color object avoidance.
Image registration is a spatial alignment of two or more images and essential technology in image... more Image registration is a spatial alignment of two or more images and essential technology in image fusion, surveillance. Unlike the electro-optics (EO) sensor, infrared (IR) sensor absorbs the radiation energy. The relation between IR and EO image of pixel intensity is sometimes similar or often reverse. To overcome this complicated problem, we propose block-based processing incorporating corner detector and descriptor. This method also uses the Gaussian pyramids to register EO/IR image. The proposed method provides more accurate registration results.
In this paper, we propose a pulmonary nodule segmentation method by using a 3D deformable model. ... more In this paper, we propose a pulmonary nodule segmentation method by using a 3D deformable model. First of all, the initial model is generated. The shape of the initial model is sphere which consists of triangle mashes. The model has internal energy and external energy. These energies are optimized to segment the nodule. Accordingly, the model is deformed through point normal vectors. After this process, the deformed models are evaluated. These processes are iterated until the model is converged.
With rapidly growing market for three-dimensional (3D) movies, 3D TVs, and 3D gaming, we may now ... more With rapidly growing market for three-dimensional (3D) movies, 3D TVs, and 3D gaming, we may now be entering a time called the "Era of 3D". The general public has also started to become comfortable with stereoscopic vision. However, there are concerns about the effects on the human body from continuously watching 3D images, for examples, visually induced motion sickness (VIMS), visual fatigue, and asthenopia. Although their mechanisms are still not fully understood, there is a great need for more knowledge about the effects of those products on users and guidelines
Abstract Three-dimensional object reconstructions is an active research area in digital imaging. ... more Abstract Three-dimensional object reconstructions is an active research area in digital imaging. In shape from focus approach, erroneous focus measurements result in inaccuracy in the depth map reconstruction of 2D object. Conventionally, to enhance the image focus volume, focus values are aggregated within a window, which is a linear filtering approach. Owing to the inherent limitation of linear process, optimal results may not be obtained. In order to overcome this limitation, a non-linear filtering approach is proposed to enhance the image focus volume for accurate depth estimation. The noisy focus values are restored in two steps. First, noisy focus values are detected using min–max operators. In order to increase the dynamic range between the minimum and the maximum focus values within the window, an appropriate power law function is designed. In second step, only the noisy measurements are replaced with the estimated ones. A refined depth map is obtained from the updated focus volume. This process continues until the difference between the previous and the current depth maps becomes very small. The performance of the proposed non-linear filtering approach is obtained for various synthetic and real objects. The results highlight the depth map estimates of the proposed approach more accurate while preserving object edges. Comparative analysis demonstrates the effectiveness of the proposed approach.
Radiologists are interested in finding the stage of cancer, so the patient can be treated and cur... more Radiologists are interested in finding the stage of cancer, so the patient can be treated and cured accordingly. This is possible by finding the type of abnormality to measure the severity of cancer in mammograms. CAD could provide them the option of better opinion about the type of abnormality. In this paper, we have proposed a novel method which can classify cancerous mammogram into six classes. Features are extracted from preprocessed images and passed through different classifiers to identify malignant mammograms and the results of winning algorithm that is Support Vector Machine (SVM) in this case are considered for next processing. Mammograms declared as malignant by SVM are divided into six classes. Again, binary classifier (SVM) is used for multi-classification using one against all technique for classification. Output of all classifiers is combined by max, median and mean rule. It has been noted that results are very much satisfactory and accuracy of classification of abnormalities is more than 96% in case of max rule. MIAS [47] data set is used for experimentation purpose.
Discrete cosine transform (DCT)-and wavelet-based schemes have been highly successful for image c... more Discrete cosine transform (DCT)-and wavelet-based schemes have been highly successful for image compression in the lossy mode. Further improvements in one or more stages of these schemes are frequently reported in literature. The input images to these methods ...
Abstract The technique to estimate the three-dimensional (3D) geometry of an object from a sequen... more Abstract The technique to estimate the three-dimensional (3D) geometry of an object from a sequence of images obtained at different focus settings is called shape from focus (SFF). In SFF, the measure of focus — sharpness — is the crucial part for final 3D shape estimation. However, it is difficult to compute accurate and precise focus value because of the noise presence during the image acquisition by imaging system. Various noise filters can be employed to tackle this problem, but they also remove the sharpness information in addition to the noise. In this paper, we propose a method based on mean shift algorithm to remove noise introduced by the imaging process while minimising loss of edges. We test the algorithm in the presence of Gaussian noise and impulse noise. Experimental results show that the proposed algorithm based on the mean shift algorithm provides better results than the traditional focus measures in the presence of the above mentioned two types of noise.
A new method for the three-dimensional shape recovery from image focus is proposed. The method is... more A new method for the three-dimensional shape recovery from image focus is proposed. The method is based on approximation of the Focussed Image Surface (FIS) by a piecewise curved surface which tracks the realistic FIS in image space. The piecewise ...
... Muhammad Bilal Ahmad and Tae-Sun Choi, SeniorMember IEEE ... [4] Muhammad Bilal Ahmad and Tae... more ... Muhammad Bilal Ahmad and Tae-Sun Choi, SeniorMember IEEE ... [4] Muhammad Bilal Ahmad and Tae-Sun Choi, A heuristic approach for finding best focused shape, IEEE Transactions on Circuit System for Video Technologies (to be published in April 2005 issue). II - 972
Abstract-When a stego image undergoes an attack, the embedded watermark is also distorted. This p... more Abstract-When a stego image undergoes an attack, the embedded watermark is also distorted. This pertains to the ability of a watermark holding to the host image and undergoing the same attack as the host image does. However, this attribute of the ...
We present a novel approach to developing Machine Learning (ML) based decoding models for extract... more We present a novel approach to developing Machine Learning (ML) based decoding models for extracting a watermark in the presence of attacks. Statistical characterization of the components of various frequency bands is exploited to allow blind extraction of the ...
... 316-317, 2001. [4] Y. K. Seong, Y. H. Choi, and T. S. Choi, Design and Implementation of Har... more ... 316-317, 2001. [4] Y. K. Seong, Y. H. Choi, and T. S. Choi, Design and Implementation of Hard Disk Drive Embedded Digital Satellite Receiver, IEEE International Symposium on Consumer Electronics, pp.168-173, 2001. [5] U. Gargi, R. Kasturi, and SH Strayer, Performance ...
This paper suggests a simple scheme in which the search pattern is divided into a number of secto... more This paper suggests a simple scheme in which the search pattern is divided into a number of sectors based on the spatio-temporal correlation information. In the first step five neighboring blocks are searched for finding the predicted motion vector. The predictive motion vector thus obtained is chosen as the initial search center. This predictive search center is found to be closer to the global minimum and thus decreases the effects of the monotonic error surface assumption and its impact on the motion field estimates. Secondly the prediction information is used to obtain the direction of predicted motion vector. Based on the direction of predicted motion vector the search area is divided into four sectors. Final search pattern is adaptive and depends on the sector selected and significantly reduces the computational complexity. Experiments show the speed improvement of the proposed algorithm as compared to other fast search algorithms, in addition the image quality measured in terms of PSNR also shows good results.
ABSTRACT This paper presents a fast algorithm for H.264 fractional motion estimation (ME). In H.2... more ABSTRACT This paper presents a fast algorithm for H.264 fractional motion estimation (ME). In H.264 ME is the most time consuming component. The ME process consists of two stages: integer pixel and fractional pixel search. To reduce the complexity of fractional pixel ME we propose a quadrant based directional fractional pixel ME algorithm that is based on the unimodal property of the fractional pixel error surface. Better computation reduction has been achieved by using this strategy. Experimental results show that as compared to fast sub-pel ME proposed in H.264, the proposed method can speedup fractional ME, with a negligible degradation in video quality.
ABSTRACT This paper suggests a simple scheme based on spatio-temporal neighborhood information fo... more ABSTRACT This paper suggests a simple scheme based on spatio-temporal neighborhood information for obtaining better estimates of motion vectors. The estimated motion vector is chosen as initial search center. This predictive search center is found closer to the global minimum. Based on the prediction, the algorithm also chooses between center-biased or uniform approach for slow or fast moving sequences. For final fine search quadrant selection algorithm is chosen that speeds up the process. Experimental results presented in this paper demonstrate the efficiency of the proposed approach.
Previous studies show that the motion vector distribution within the search window shows a center... more Previous studies show that the motion vector distribution within the search window shows a center biased behavior. Based on this fact, an advanced center biased search algorithm for fast block motion estimation has been proposed. The algorithm drastically reduces the computational complexity by strict application of the unimodal error surface assumption. Improved error performance has been achieved by an efficient center biased search strategy that improves the chances of getting the correct motion vector. The half stop technique has been adopted to speed up the block matching process. Experimental results show that proposed algorithm has improved performance as compared to the three step search algorithm. A good predicted image quality is also achieved
Uploads
Papers by Tae-sun Choi