Rahil Garnavi - Academia.edu (original) (raw)

Papers by Rahil Garnavi

Research paper thumbnail of Classification of Melanoma Lesions Using Wavelet-Based Texture Analysis

... optimal feature subsets, four different classifiers of support vector machine, random forest,... more ... optimal feature subsets, four different classifiers of support vector machine, random forest, logistic model tree and hidden naive bayes, have been applied on a test set of 102 dermoscopy images. ... 157, pp. 926–933, 2007. [7] M. Celebi, H. Kingravi, B. Uddin, H. Iyatomi, Y ...

Research paper thumbnail of Border detection in dermoscopy images using hybrid thresholding on optimized color channels

Computerized Medical Imaging and Graphics, 2011

Automated border detection is one of the most important steps in dermoscopy image analysis. Altho... more Automated border detection is one of the most important steps in dermoscopy image analysis. Although numerous border detection methods have been developed, few studies have focused on determining the optimal color channels for border detection in dermoscopy images. This paper proposes an automatic border detection method which determines the optimal color channels and performs hybrid thresholding to detect the lesion borders. The color optimization process is tested on a set of 30 dermoscopy images with four sets of dermatologist-drawn borders used as the ground truth. The hybrid border detection method is tested on a set of 85 dermoscopy images with two sets of ground truth using various metrics including accuracy, precision, sensitivity, specificity, and border error. The proposed method, which is comprised of two stages, is designed to increase specificity in the first stage and sensitivity in the second stage. It is shown to be highly competitive with three state-of-the-art border detection methods and potentially faster, since it mainly involves scalar processing as opposed to vector processing performed in the other methods. Furthermore, it is shown that our method is as good as, and in some cases more effective than a dermatology registrar.

Research paper thumbnail of Automatic Segmentation of Dermoscopy Images Using Histogram Thresholding on Optimal Color Channels

Automatic segmentation of skin lesions is the first step towards development of a computer-aided ... more Automatic segmentation of skin lesions is the first step towards development of a computer-aided diagnosis of melanoma. Although numerous segmentation methods have been developed, few studies have focused on determining the most discriminative and effective color space for melanoma application. This paper proposes a novel automatic segmentation algorithm using color space analysis and clustering-based histogram thresholding, which is able to determine the optimal color channel for segmentation of skin lesions. To demonstrate the validity of the algorithm, it is tested on a set of 30 high resolution dermoscopy images and a comprehensive evaluation of the results is provided, where borders manually drawn by four dermatologists, are compared to automated borders detected by the proposed algorithm. The evaluation is carried out by applying three previously used metrics of accuracy, sensitivity, and specificity and a new metric of similarity. Through ROC analysis and ranking the metrics, it is shown that the best results are obtained with the X and XoYoR color channels which results in an accuracy of approximately 97%. The proposed method is also compared with two state-ofthe-art skin lesion segmentation methods, which demonstrates the effectiveness and superiority of the proposed segmentation method.

Research paper thumbnail of Global versus Hybrid Thresholding for Border Detection in Dermoscopy Images

In this paper we demonstrate the superiority of the automated hybrid thresholding approach to bor... more In this paper we demonstrate the superiority of the automated hybrid thresholding approach to border detection in dermoscopy images over the global thresholding method through a newly introduced evaluation metric: Performance Index. The approach incorporates optimal color channels into the hybrid thresholding method, which is a combination of global and adaptive local thresholding, to determine the closest border to that drawn by dermatologists. Statistical analysis and optimization procedure are used and shown to be convergent in determining the optimal parameters for the local thresholding procedure in order to obtain the most accurate borders. The effectiveness of the approach is tested on 55 high resolution dermoscopy images of patients, with manual borders drawn by three expert dermatologists, and the union is used as the ground truth. The results demonstrate the significant advantages of the automated hybrid approach over the global thresholding method.

Research paper thumbnail of Optimized Weighted Performance Index for Objective Evaluation of Border-Detection Methods in Dermoscopy Images

IEEE Transactions on Information Technology in Biomedicine, 2011

Quantitative evaluation of the existing border-detection methods is commonly performed by using d... more Quantitative evaluation of the existing border-detection methods is commonly performed by using different metrics. This is inherently problematic due to the different characteristics of each metric. This paper presents a novel approach for objective evaluation of border-detection methods in dermoscopy images by introducing a comprehensive evaluation metric: optimized weighted performance index. The index is formed as a nonlinear weighted function of the six commonly used metrics of sensitivity, specificity, accuracy, precision, border error, and similarity. Constrained nonlinear multivariable optimization techniques are applied to determine the optimal set of weights that result in the maximum value of the index. This index is used as an effective measure of the value of a given border-detection method and, thus, provides a basis for comparison with other methods. To demonstrate the effectiveness of the proposed index, it is used to evaluate five recent border-detection methods applied on a set of 55 high-resolution dermoscopy images.

Research paper thumbnail of Weighted performance index for objective evaluation of border detection methods in dermoscopy images: Evaluation of border detection via WPI

Skin Research and Technology, 2010

Purpose: This paper presents a novel approach for objective evaluation of border detection in der... more Purpose: This paper presents a novel approach for objective evaluation of border detection in dermoscopy images of melanoma.Background: In melanoma studies, border detection is a fundamental step toward the development of a computer-aided diagnosis system. Therefore, its accuracy is essential for accurate implementation of the subsequent parts of the diagnostic system.Method: An objective evaluation procedure of border detection methods is presented. The evaluation procedure uses the weighted performance index, which is composed of weighted metrics of sensitivity, specificity, accuracy, precision, border error and similarity. This index can also be used to optimize the parameters of a border detection method.Result and conclusion: Experiments are performed on 55 high-resolution dermoscopy images. Using the union of four sets of dermatologist-drawn borders as the ground truth, weighted metrics of sensitivity, specificity, accuracy, precision, border error and similarity are evaluated. Then, the weighted performance index is constructed and used to optimize the parameters of the hybrid border detection method. The outcome of the optimization process, verified through statistical analysis, yields a higher degree of agreement between automatic borders and the ground truth, compared with using standard metrics only. Finally, the weighted performance index is used to evaluate five recently reported border detection methods.

Research paper thumbnail of Skin Lesion Segmentation Using Color Channel Optimization and Clustering-based Histogram Thresholding

Abstract—Automatic segmentation of skin lesions is the first step towards the automated analysis ... more Abstract—Automatic segmentation of skin lesions is the first step towards the automated analysis of malignant melanoma. Although numerous segmentation methods have been developed, few studies have focused on determining the most effective color space for melanoma ...

Research paper thumbnail of Image based diagnostic aid system for interstitial lung diseases

Expert Systems With Applications, 2011

Automatic classification of lung tissue patterns in high-resolution computed tomography (HRCT) im... more Automatic classification of lung tissue patterns in high-resolution computed tomography (HRCT) images of patients affected with interstitial lung diseases (ILD) is an important stage in the construction of a computer-aided diagnosis system. In this study, we propose a new image based system for classification of lung tissue patterns. The proposed system comprises three stages. In the first stage, the parenchyma region in HRCT lung images is separated using a set of thresholding, filtering and morphological operators. In the second stage, two sets of overcomplete wavelet filters, namely discrete wavelet frames and rotated wavelet frames, are utilized to extract features from defined regions of interest (ROIs) within parenchyma. Then, in the third stage, the fuzzy k-nearest neighbor algorithm is employed to perform the pattern classification. The proposed method is tested for classifying four different lung tissue patterns (ground glass, honeycombing, reticular, and normal) selected from a database of 339 images from 17 subjects. After applying our technique to classify these patterns in isolated ROIs, we extend the classification scheme to the whole lung in order to produce quantitative scores of abnormalities in lung parenchyma of patients. The performance of the proposed method is compared with two state-of-the-art texture based methods for lung tissue characterization and is also validated against experienced observers. The average kappa statistic of the agreement between two radiologists and the computer was found to be 0.6543 where as the average kappa statistic for the inter-observer agreement was 0.6848. We also performed an experiment to show the correlation between pulmonary function test parameters and quantitative scores of computerized system. Results show that extent of HRCT findings correlates significantly with functional impairment. The computer system is shown to approach the performance of the expert observers in diagnosing regions of interest and can help to produce objective measures of abnormal patterns in lung HRCT images.

Research paper thumbnail of A New Segmentation Method for Lung HRCT Images

Image segmentation plays a crucial role in many medical imaging applications by automating or fac... more Image segmentation plays a crucial role in many medical imaging applications by automating or facilitating the delineation of anatomical structures and other regions of interest. The aim of this paper is to develop an accurate and reliable method for segmentation of lung HRCT images using a pixelbased approach.

Research paper thumbnail of Texture Analysis in Lung HRCT Images

Automatic classification of lung tissue patterns in high resolution computed tomography images of... more Automatic classification of lung tissue patterns in high resolution computed tomography images of patients with interstitial lung diseases is an important stage in the construction of a computer-aided diagnosis system. To this end, a novel approach is proposed using two sets of overcomplete wavelet filters, namely discrete wavelet frames (DWF) and rotated wavelet frames (RWF), to extract the features which best characterizes the lung tissue patterns. Support vector machines learning algorithm is then applied to perform the pattern classification. Four different lung patterns (ground glass, honey combing, reticular, and normal) selected from a database of 340 images are classified using the proposed method. The overall multiclass accuracy reaches 90. 72%, 95.85%, and 96.81% for DWF, RWF, and their combination, respectively. These results prove that RWF is superior to DWF, due to its orientation selectivity. However, best results are obtained by the combination of two filter banks which shows that the two set of filters are complementary.

Research paper thumbnail of Classification of Melanoma Lesions Using Wavelet-Based Texture Analysis

... optimal feature subsets, four different classifiers of support vector machine, random forest,... more ... optimal feature subsets, four different classifiers of support vector machine, random forest, logistic model tree and hidden naive bayes, have been applied on a test set of 102 dermoscopy images. ... 157, pp. 926–933, 2007. [7] M. Celebi, H. Kingravi, B. Uddin, H. Iyatomi, Y ...

Research paper thumbnail of Border detection in dermoscopy images using hybrid thresholding on optimized color channels

Computerized Medical Imaging and Graphics, 2011

Automated border detection is one of the most important steps in dermoscopy image analysis. Altho... more Automated border detection is one of the most important steps in dermoscopy image analysis. Although numerous border detection methods have been developed, few studies have focused on determining the optimal color channels for border detection in dermoscopy images. This paper proposes an automatic border detection method which determines the optimal color channels and performs hybrid thresholding to detect the lesion borders. The color optimization process is tested on a set of 30 dermoscopy images with four sets of dermatologist-drawn borders used as the ground truth. The hybrid border detection method is tested on a set of 85 dermoscopy images with two sets of ground truth using various metrics including accuracy, precision, sensitivity, specificity, and border error. The proposed method, which is comprised of two stages, is designed to increase specificity in the first stage and sensitivity in the second stage. It is shown to be highly competitive with three state-of-the-art border detection methods and potentially faster, since it mainly involves scalar processing as opposed to vector processing performed in the other methods. Furthermore, it is shown that our method is as good as, and in some cases more effective than a dermatology registrar.

Research paper thumbnail of Automatic Segmentation of Dermoscopy Images Using Histogram Thresholding on Optimal Color Channels

Automatic segmentation of skin lesions is the first step towards development of a computer-aided ... more Automatic segmentation of skin lesions is the first step towards development of a computer-aided diagnosis of melanoma. Although numerous segmentation methods have been developed, few studies have focused on determining the most discriminative and effective color space for melanoma application. This paper proposes a novel automatic segmentation algorithm using color space analysis and clustering-based histogram thresholding, which is able to determine the optimal color channel for segmentation of skin lesions. To demonstrate the validity of the algorithm, it is tested on a set of 30 high resolution dermoscopy images and a comprehensive evaluation of the results is provided, where borders manually drawn by four dermatologists, are compared to automated borders detected by the proposed algorithm. The evaluation is carried out by applying three previously used metrics of accuracy, sensitivity, and specificity and a new metric of similarity. Through ROC analysis and ranking the metrics, it is shown that the best results are obtained with the X and XoYoR color channels which results in an accuracy of approximately 97%. The proposed method is also compared with two state-ofthe-art skin lesion segmentation methods, which demonstrates the effectiveness and superiority of the proposed segmentation method.

Research paper thumbnail of Global versus Hybrid Thresholding for Border Detection in Dermoscopy Images

In this paper we demonstrate the superiority of the automated hybrid thresholding approach to bor... more In this paper we demonstrate the superiority of the automated hybrid thresholding approach to border detection in dermoscopy images over the global thresholding method through a newly introduced evaluation metric: Performance Index. The approach incorporates optimal color channels into the hybrid thresholding method, which is a combination of global and adaptive local thresholding, to determine the closest border to that drawn by dermatologists. Statistical analysis and optimization procedure are used and shown to be convergent in determining the optimal parameters for the local thresholding procedure in order to obtain the most accurate borders. The effectiveness of the approach is tested on 55 high resolution dermoscopy images of patients, with manual borders drawn by three expert dermatologists, and the union is used as the ground truth. The results demonstrate the significant advantages of the automated hybrid approach over the global thresholding method.

Research paper thumbnail of Optimized Weighted Performance Index for Objective Evaluation of Border-Detection Methods in Dermoscopy Images

IEEE Transactions on Information Technology in Biomedicine, 2011

Quantitative evaluation of the existing border-detection methods is commonly performed by using d... more Quantitative evaluation of the existing border-detection methods is commonly performed by using different metrics. This is inherently problematic due to the different characteristics of each metric. This paper presents a novel approach for objective evaluation of border-detection methods in dermoscopy images by introducing a comprehensive evaluation metric: optimized weighted performance index. The index is formed as a nonlinear weighted function of the six commonly used metrics of sensitivity, specificity, accuracy, precision, border error, and similarity. Constrained nonlinear multivariable optimization techniques are applied to determine the optimal set of weights that result in the maximum value of the index. This index is used as an effective measure of the value of a given border-detection method and, thus, provides a basis for comparison with other methods. To demonstrate the effectiveness of the proposed index, it is used to evaluate five recent border-detection methods applied on a set of 55 high-resolution dermoscopy images.

Research paper thumbnail of Weighted performance index for objective evaluation of border detection methods in dermoscopy images: Evaluation of border detection via WPI

Skin Research and Technology, 2010

Purpose: This paper presents a novel approach for objective evaluation of border detection in der... more Purpose: This paper presents a novel approach for objective evaluation of border detection in dermoscopy images of melanoma.Background: In melanoma studies, border detection is a fundamental step toward the development of a computer-aided diagnosis system. Therefore, its accuracy is essential for accurate implementation of the subsequent parts of the diagnostic system.Method: An objective evaluation procedure of border detection methods is presented. The evaluation procedure uses the weighted performance index, which is composed of weighted metrics of sensitivity, specificity, accuracy, precision, border error and similarity. This index can also be used to optimize the parameters of a border detection method.Result and conclusion: Experiments are performed on 55 high-resolution dermoscopy images. Using the union of four sets of dermatologist-drawn borders as the ground truth, weighted metrics of sensitivity, specificity, accuracy, precision, border error and similarity are evaluated. Then, the weighted performance index is constructed and used to optimize the parameters of the hybrid border detection method. The outcome of the optimization process, verified through statistical analysis, yields a higher degree of agreement between automatic borders and the ground truth, compared with using standard metrics only. Finally, the weighted performance index is used to evaluate five recently reported border detection methods.

Research paper thumbnail of Skin Lesion Segmentation Using Color Channel Optimization and Clustering-based Histogram Thresholding

Abstract—Automatic segmentation of skin lesions is the first step towards the automated analysis ... more Abstract—Automatic segmentation of skin lesions is the first step towards the automated analysis of malignant melanoma. Although numerous segmentation methods have been developed, few studies have focused on determining the most effective color space for melanoma ...

Research paper thumbnail of Image based diagnostic aid system for interstitial lung diseases

Expert Systems With Applications, 2011

Automatic classification of lung tissue patterns in high-resolution computed tomography (HRCT) im... more Automatic classification of lung tissue patterns in high-resolution computed tomography (HRCT) images of patients affected with interstitial lung diseases (ILD) is an important stage in the construction of a computer-aided diagnosis system. In this study, we propose a new image based system for classification of lung tissue patterns. The proposed system comprises three stages. In the first stage, the parenchyma region in HRCT lung images is separated using a set of thresholding, filtering and morphological operators. In the second stage, two sets of overcomplete wavelet filters, namely discrete wavelet frames and rotated wavelet frames, are utilized to extract features from defined regions of interest (ROIs) within parenchyma. Then, in the third stage, the fuzzy k-nearest neighbor algorithm is employed to perform the pattern classification. The proposed method is tested for classifying four different lung tissue patterns (ground glass, honeycombing, reticular, and normal) selected from a database of 339 images from 17 subjects. After applying our technique to classify these patterns in isolated ROIs, we extend the classification scheme to the whole lung in order to produce quantitative scores of abnormalities in lung parenchyma of patients. The performance of the proposed method is compared with two state-of-the-art texture based methods for lung tissue characterization and is also validated against experienced observers. The average kappa statistic of the agreement between two radiologists and the computer was found to be 0.6543 where as the average kappa statistic for the inter-observer agreement was 0.6848. We also performed an experiment to show the correlation between pulmonary function test parameters and quantitative scores of computerized system. Results show that extent of HRCT findings correlates significantly with functional impairment. The computer system is shown to approach the performance of the expert observers in diagnosing regions of interest and can help to produce objective measures of abnormal patterns in lung HRCT images.

Research paper thumbnail of A New Segmentation Method for Lung HRCT Images

Image segmentation plays a crucial role in many medical imaging applications by automating or fac... more Image segmentation plays a crucial role in many medical imaging applications by automating or facilitating the delineation of anatomical structures and other regions of interest. The aim of this paper is to develop an accurate and reliable method for segmentation of lung HRCT images using a pixelbased approach.

Research paper thumbnail of Texture Analysis in Lung HRCT Images

Automatic classification of lung tissue patterns in high resolution computed tomography images of... more Automatic classification of lung tissue patterns in high resolution computed tomography images of patients with interstitial lung diseases is an important stage in the construction of a computer-aided diagnosis system. To this end, a novel approach is proposed using two sets of overcomplete wavelet filters, namely discrete wavelet frames (DWF) and rotated wavelet frames (RWF), to extract the features which best characterizes the lung tissue patterns. Support vector machines learning algorithm is then applied to perform the pattern classification. Four different lung patterns (ground glass, honey combing, reticular, and normal) selected from a database of 340 images are classified using the proposed method. The overall multiclass accuracy reaches 90. 72%, 95.85%, and 96.81% for DWF, RWF, and their combination, respectively. These results prove that RWF is superior to DWF, due to its orientation selectivity. However, best results are obtained by the combination of two filter banks which shows that the two set of filters are complementary.