Ultrasound Kidney Images Research Articles

A Dynamic Graph Cuts Method with Integrated Multiple Feature Maps for Segmenting Kidneys in 2D Ultrasound Images

Automatic segmentation of kidneys in ultrasound (US) images remains a challenging task because of high speckle noise, low contrast, and large appearance variations of kidneys in US images. Because texture features may improve the US image segmentation performance, we propose a novel graph cuts method to segment kidney in US images by integrating image intensity information and texture feature maps. We develop a new graph cuts-based method to segment kidney US images by integrating original image intensity information and texture feature maps extracted using Gabor filters. To handle large appearance variation within kidney images and improve computational efficiency, we build a graph of image pixels close to kidney boundary instead of building a graph of the whole image. To make the kidney segmentation robust to weak boundaries, we adopt localized regional information to measure similarity between image pixels for computing edge weights to build the graph of image pixels. The localized graph is dynamically updated and the graph cuts-based segmentation iteratively progresses until convergence. Our method has been evaluated based on kidney US images of 85 subjects. The imaging data of 20 randomly selected subjects were used as training data to tune parameters of the image segmentation method, and the remaining data were used as testing data for validation. Experiment results demonstrated that the proposed method obtained promising segmentation results for bilateral kidneys (average Dice index = 0.9446, average mean distance = 2.2551, average specificity = 0.9971, average accuracy = 0.9919), better than other methods under comparison (P < .05, paired Wilcoxon rank sum tests). The proposed method achieved promising performance for segmenting kidneys in two-dimensional US images, better than segmentation methods built on any single channel of image information. This method will facilitate extraction of kidney characteristics that may predict important clinical outcomes such as progression of chronic kidney disease.

Analysis of polycystic kidney disease in medical ultrasound images

The growth of kidney diseases has gradually increased in recent years. Ultrasound imaging provides the internal structure of the body to detect eventually diseases or abnormal tissues non-invasively. Segmentation of required region in ultrasound images is one of the challenging tasks. The proposed method focuses on classification of medical ultrasound images of kidney as cystic and polycystic types. Segmentation is performed using gradient vector force (GVF) snakes. Before segmentation, speckle noise is removed using Gaussian filter and contrast is enhanced. We have segmented normal, cystic and polycystic kidney ultrasound images effectively using GVF snakes. We have also carried out segmentation using morphological operations which requires a user intervention during the process of segmentation. Geometrical features are used with k-NN for classifying medical US images of kidney as normal, single cystic and polycystic types for segmented regions .The proposed method has applications in analysis of organ morphology and realising quantitative measurements.

Efficient Segmentation of ultrasound images of abnormal kidney

Image processing has large applications in the medical diagnosing, especially in Ultrasound Imaging (UI) that for its' safety. UI suffer from low resolution and noises, so, became a target for a lot of investigations. Three US images which represent Cyst, Benign, and Stone used in this study. Five contrast enhancement techniques and three denoising filter applied on them and best of them chose according to their PSNR and MSE values. It's found that best contrast enhancement methods were Contrast adjustment, for Benign and Cyst images, and sharpening for the Stone image. Then, the images denoised by Median filtering due to found it as better denoising filter than others studied here. Three segmenting methods (one ordinary and two hybrid techniques) were used here: thresholding, kmean with intensity selector, and bilateral filter with intensity selector. All techniques were segmented nearly same area of the cases. But thresholding needs a large of trial and error to obtain their segmenting. The kmean with intensity selector method was found better for the Benign case than other for its higher PSNR (16.68), lower MSE (1394.1), and small time of running (22.1 sec). But for Cyst and Stone images, Bilateral filter with intensity selector was found better than other according to same parameters. So, techniques of hybrid segmentation provided more efficient segmenting.

Automatic renal lesion segmentation in ultrasound images based on saliency features, improved LBP, and an edge indicator under level set framework.

Segmentation of lesions in ultrasound images is widely used for preliminary diagnosis. In this paper, we develop an automatic segmentation algorithm for multiple types of lesions in ultrasound images. The proposed method is able to detect and segment lesions automatically as well as generate accurate segmentation results for lesion regions. In the detection step, two saliency detection frameworks which adopt global image information are designed to capture the differences between normal and abnormal organs as well as these between lesions and the normal tissues around them. In the segmentation step, three types of local information, i.e., image intensity, improved local binary patterns (LBP) features, and an edge indicator, are embedded into a modified level set framework to carry out the segmentation task. The cyst and carcinoma regions in the ultrasound images of the human kidneys can be automatically detected and segmented by using the proposed method. The efficiency and accuracy of the method are validated by quantitative evaluations and comparative measurements with three well-recognized segmentation methods. Specifically, the average precision and dice coefficient of the proposed method in segmenting renal cysts are 95.33% and 90.16%, respectively, while those in segmenting renal carcinomas are 94.22% and 91.13%, respectively. The average precision and dice coefficient of the proposed method are higher than those of three compared segmentation methods. The proposed method can efficiently detect and segment the renal lesions in ultrasound images. In addition, since the proposed method utilizes the differences between normal and abnormal organs as well as these between lesions and the normal tissues around them, it can be possibly extended to deal with lesions in other organs of ultrasound images as well as lesions in medical images of other modalities.

Texture Feature Analysis for Different Resolution Level of Kidney Ultrasound Images

Image feature extraction is a technique to identify the characteristic of the image. The objective of this work is to discover the texture features that best describe a tissue characteristic of a healthy kidney from ultrasound (US) image. Three ultrasound machines that have different specifications are used in order to get a different quality (different resolution) of the image. Initially, the acquired images are pre-processed to de-noise the speckle to ensure the image preserve the pixels in a region of interest (ROI) for further extraction. Gaussian Low- pass Filter is chosen as the filtering method in this work. 150 of enhanced images then are segmented by creating a foreground and background of image where the mask is created to eliminate some unwanted intensity values. Statistical based texture features method is used namely Intensity Histogram (IH), Gray-Level Co-Occurance Matrix (GLCM) and Gray-level run-length matrix (GLRLM).This method is depends on the spatial distribution of intensity values or gray levels in the kidney region. By using One-Way ANOVA in SPSS, the result indicated that three features (Contrast, Difference Variance and Inverse Difference Moment Normalized) from GLCM are not statistically significant; this concludes that these three features describe a healthy kidney characteristics regardless of the ultrasound image quality.

An Approach towards Ultrasound Kidney Cysts Detection using Vector Graphic Image Analysis

This study develops new approach towards detection of kidney ultrasound image for both with single cyst as well as multiple cysts. 50 single cyst images and 25 multiple cysts images were used to test the developed algorithm. Steps involved in developing this algorithm were vector graphic image formation and analysis, thresholding, binarization, filtering as well as roundness test. Performance evaluation to 50 single cyst images gave accuracy of 92%, while for multiple cysts images, the accuracy was about 86.89% when tested to 25 multiple cysts images. This developed algorithm may be used in developing a computerized system such as computer aided diagnosis system to help medical experts in diagnosis of kidney cysts.

Case 1: The Inconsolable Infant in the NICU

A female infant is born at 30 weeks of gestation via vaginal delivery due to imminent labor. The infant receives 1 dose of surfactant in the delivery room for respiratory distress, after which she is transferred to the NICU for further care. Anthropometric measurements at birth are: weight, 595 g; length, 29 cm; and head circumference, 23.5 cm. Birthweight, length, and head circumference are below the 5th percentile for the corresponding gestational age. Wide forehead, 2 natal teeth on the mandibular ridge, hypertonia of all extremities, grade 2 systolic murmur, and bilateral inguinal hernias are noted on examination. Specific immunoglobulin M values for toxoplasma, rubella and cytomegalovirus virus are negative. Head ultrasonography reveals no calcifications or other abnormalities. Skeletal survey shows no signs of perichondritis. Retinal examination to check for signs of chorioretinal inflammation that could point toward a particular TORCH (toxoplasmosis, other [syphilis, varicella-zoster, parvovirus B19], rubella, cytomegalovirus, and herpes) infection are negative. Peripheral pulmonary arteries are found to be mildly hypoplastic on echocardiography, which is conducted as part of an evaluation for an audible murmur. Ultrasound imaging of kidneys and bladder reveals no anatomic abnormalities. The patient undergoes elective extubation 24 hours after birth. She makes a transition to room air within 72 hours after birth. After extubation, she has a …

Advanced Methods for Detecting PUJ Dilatation by 2D Ultrasound

Pelviureteric junction (PUJ) obstruction is a condition frequently encountered in both adult and pediatric patients. Congenital abnormalities and crossing lower-pole renal vessels are the most common underlying pathologies in both men and women. There are different methods for detecting it the most usual, safe, and easy one is by ultrasound scanning. The aim of this study is how to improve the image quality of two dimensional (2D) ultrasound screening of detecting PUJ dilatation using image processing software, image enhancement, and different types of filters, then making comparison which filter is the best to improve the image quality, that helps the medical doctors, and sonographers to make the correct decision and diagnosis. 1357 patients scanned by ultrasound in Harer general hospital for general abdominal scanning, 987 cases among them have detected as urinary tract infection cases among this 987 case there were 73 case of them with PUJ dilatation. The 2D ultrasound images saved, after making image enhancement and using different types of filters (HE, ADH, CLAHE, and Wiener) to enhance four 2D ultrasound images of abnormal kidneys, the result was in each type of filters there were some advantages and disadvantages, so that the best type of filters are (HE and ADH) because the PUJ and pelvis is much more clear and easy to define after using these kinds of filters. Index Terms: ADH, CLAHE, HE, Image Enhancement, Image Processing, Pelviureteric Junction, Wiener

Extrapulmonary Tuberculosis in Children (Ultrasound Finding)

To describe ultrasound of extrapulmonary tuberculosis in children. This descriptive study conducted at very low economy and low hygiene sanitation environment at rural boundary islands: Saumlaki Health Centre, Tanimbar, Moluccas (January, 20-23. 2009 and March - May 2010), St Elisabeth Hospital, Flores, Nusa Tenggara Timur (February 2012 - August 2013) and Stela Maris, Lukas Hilisimetano Hospitals and Fanayama Primary Health Centret, Onohondro Village at South Nias–North Sumatera province (April- July 2015). Data were abstracted from medical records of children disseminated tuberculosis. Patients were diagnosed by one certified ultrasound internal medicine specialist using the only one black and white portable ultrasound available. Patients had never previously received medical specialist service nor ultrasound examination. N=724 children , 0,5-15y (Saumlaki 392, Flores 184 and South Nias 148). All (100%) were Peritoneal (dry type) with recurrent colic abdomen pain, diarrhea/obstipation, distention, doughy & dam-board phenomena as well as constitutional symptom and undernourished. Small bowel ultrasound showed a/hypo- peristaltic, irregular thickening heterogenic hypo-echoic, irregular margin of the wall, loss differentiation of the wall layers, with several oval/round nodular structures those suggestive granuloma. 100% Lymph node enlargement both neck and paraaortal. Kidney ultrasound images according to the kidney during childhood growth. Normal pancreas and urinary bladder. At South Nias: Musculoskeletal manifestation were 2 (1,4%). Peritoneal dry type and small bowel tuberculosis and lymph node enlargement were the most extrapulmonary tuberculosis, followed by musculoskeletal.

Renal Allograft Aspergillus Infection Presenting With Obstructive Uropathy: A Case Report

BackgroundIsolated renal allograft aspergillosis is rare and usually presents with fever and decreased glomerular filtration rate. Presentation with obstructive uropathy caused by aspergillus fungal balls is much less common. We report a young male patient who presented with obstructive uropathy secondary to isolated renal allograft aspergillus infection 6 weeks after transplant. He was treated with nephrectomy and antifungal medications. Case presentationA 29-year-old Saudi male patient had a recent living non-related kidney transplantation in Pakistan. Early Post-transplant course was complicated by acute cellular rejection (Banff Class IB) which was managed successfully with pulse steroid and anti-thymocyte globulin. The patient presented again to our emergency room on fortieth day post-transplant with a complaint of decreased urine output and passing white particles in his urine. This presentation was three Three weeks after treatment for cellular rejection, the urine fungal culture showed growth of Aspergillus fumigatus, and ultrasound imaging of the allograft kidney revealed mild to moderate hydronephrosis with echogenic materials within the renal pelvis. Biopsy of the transplanted kidney showed severe necrotizing granulomatous inflammation and fungal elements consistent with aspergillus species. The patient was given voriconazole as an antifungal agent and was weaned from immunosuppressive medication. The patient eventually required intermittent hemodialysis and underwent surgical allograft nephrectomy. ConclusionSuboptimal environmental and infection prevention and control precautions can explain this type of infection. It is important for clinicians to have a high index of suspicion and to investigate for fungal infection as a rare cause of obstructive uropathy in high-risk patients.

Phase based distance regularized level set for the segmentation of ultrasound kidney images

Segmentation of ultrasound images has been considered as a challenging task due to its low contrast characteristics. Expert's manual segmentation not only relies on just intensity information available with the image because human vision system roughly relies on intensity but at the same time, it reads local information too. Hence, an attempt has been made, where local properties such as phase, energy and amplitude were extracted from the gray scale image using monogenic signals. Here, the monogenic signals were constructed using Cauchy functions rather than traditional Log Gabor function. Active contour based segmentation that incorporates distance regularized level set evolution concept has been used to segment the region of interest on the local information based image i.e Phase Based Distance Regularized Level Set. Proposed method utilizes local phase and feature asymmetry in achieving the objective. Robustness of the proposed method is demonstrated with the quantitative metrics such as Dice Similarity Coefficient (DSC) index, Hausdorff Distance (HD) and Mean Absolute Deviation (MAD) measures. Proposed method is compared quantitatively with the intensity based Distance Regularized Level Set Evolution (DRLSE), with Geodesic Active Contour (GAC) and with the variant of the proposed method that uses Log Gabor filter bank instead of Cauchy.

Comparative estimation of the diagnostic value of radiographic and ultrasound studies in microbial inflammatory diseases of the urinary tract in children

The results of using three imaging technologies (intravenous urography, micturating cystography, and B-mode kidney ultrasound imaging) in 62 children aged 10 to 206 months, examined at the city nephrology hospital, were comparatively analyzed. The findings showed the comparatively high diagnostic value (a sensitivity of 70% and a specificity of 87%) of echography particularly for severe urinary tract malformations. At the same time, echo location was found to be insufficient to detect vesicoureteral reflux (a sensitivity of 20% and a specificity of 85,5%), which is a ground for the further improvement of examination protocols for patients with urinary tract infections.

A Technique to Suppress Speckle in Ultrasound Images using Nonlocal Mean and Cellular Automata

Kidney ultrasound imaging is an economical, non invasive, real time diagnosis system which is used to measure abnormalities in the shape, size and location of the kidneys in the body. The presence of speckle noise degrades the visual quality of ultrasound images. The primary objective of the research under taken is to detect the renal stone in ultrasound images. The present research paper discusses an algorithm based on cellular automata and nonlocal mean for suppression of the speckle noise. The cellular automatons are dynamic systems to represent the particular problem in terms of specific pattern. This concept is exploited to distinguish the noise from the object being target. The kidney ultrasound images are denoised to detect presence of renal stone in the image. The concept behind the nonlocal mean is an assumption that the pixel being considered for denosing has strong connection with the surrounding area rather than the exactly nearly by pixels. The MLNC (Maximum Likelihood Neighborhood Computation) rules are applied to find the nearest value in the neighbour. The BEP (Border, Edge Preservation) rules are used to safeguard the edges between the regions. The samples of 55 clinical images of ultrasound are used for statistical analysis which is carried out by Signal to Noise Ratio (SNR), Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR). The result analysis shows MSE = 216.789, SNR = 49.1569 and PSNR = 68.022. The algorithm not only enhances the visual quality but also suppress unwanted speckle noise. The algorithm is dynamic in nature and can be used to de-noise SAR images, MRI and X-rays image. The method is adaptive in nature as cellular modeling can model low noise level.

A Multiresolution Bilateral Filter for Speckle Reduction in Ultrasound Kidney Images

The kidney stones (renal calculi) are the most common medical problems that occur to the people, irrespective of age group. The ultrasonography is a versatile and noninvasive imaging modality often preferred over other techniques such as CT, MRI and X-ray for kidney stone detection. It does not use any ionizing radiation for its operation. However, the main disadvantage associated with ultrasound scanning is the presence of speckle noise which mimics as a kidney stone and leads to fault diagnosis. The existence of speckle affects the further image processing steps such as feature extraction, segmentation, and classification of stones. A new wavelet based bilateral filtering technique is proposed for speckle reduction of ultrasound kidney images in order to make the diagnosis procedure of kidney stones in the easiest way. In this method, the approximation coefficients of decomposed image are processed using the bilateral filter. The noise reduction in detail coefficient is performed using wavelet soft thresholding technique. The threshold value is derived from a Bayesian framework. This method of speckle reduction is applied on various ultrasound kidney images with stone and visual investigation made on the filter performance in terms of the identification of actual stones by reducing the effect of speckle resembles like a stone. Also, the effectiveness of the proposed method is examined by means of RMSE, PSNR, and SSIM. The proposed hybrid architecture is compared with wavelet Hard and Soft thresholding and Bilateral Filter.

A Robust Edge Preserving Bilateral Filter for Ultrasound Kidney Image

Objective: The speckle noise reduction is an important preprocessing step, normally performed to improve the segmentation of kidney images for the diagnosis of stones' presence or its size. The bilateral filter is a robust method for noise reduction with edge preservation. An attempt is made in this work to analyze the effectiveness of the bilateral filter for speckle reduction in ultrasound kidney images. Methods/Statistical Analysis: An open source software Fiji is used to develop the filtering algorithms for ultrasound kidney image. The bilateral filter performance is compared with familiar speckle filters named median filter, Speckle Reducing Anisotropic Diffusion (SRAD) filter and Non-Local Mean (NLM) filter. The authors have selected mainly Root Mean Square Error (RMSE) for image quality analysis, Signal to Noise Ratio (SNR) for comparison of noise with useful signal strength, Peak Signal to Noise Ratio (PSNR) for maximum amplitude of signal, Mean Absolute Error (MAE) for overall performance and Structural Similarity Index Measure (SSIM) for testing bilateral filter on kidney images. The statistical measures also calculated to distinguish the filter performances. Findings: The bilateral filter performance is proven as an effective method for speckle reduction in ultrasound kidney stone detection applications than the existing methods through quantitative and statistical analysis. Application/Improvement: The bilateral filter can offer better segmentation of kidney stone due to its effective speckle reduction.

Development of Patient Assistive Tool for Detection of Abnormalities in Kidney

Ultrasound Scan is the most popular and cheapest imaging technique which is used as a preliminary investigating tool by the doctors nowadays. In this regard, development of a computer aided assistive tool is utmost necessary to create a user friendly interface for the patients to expose them to basic treatment procedures available in case of abnormalities found in the Ultrasound Scan. An easy and better understanding of the impressions given in the scan is vital for further treatment procedures. As a first step towards creating a complete computer aided treatment assistive tool, the abnormalities detection in kidney ultrasound images is taken up as a primary goal in this paper. The most common renal problems that can be identified in the ultrasound images are renal stones or calculi, cyst and infections. A Graphical User Interface (GUI) is developed to display the count and size of calculi present in kidney. The GUI also specifies the presence of cyst. Ultrasound renal images are preprocessed and the speckle noises are removed using median filter. MATLAB software is used to perform the image processing. Segmentation is performed using Thresholding and Seeded Region growing algorithms. The algorithms are compared based on the statistical features which are extracted from the segmented images and the Seeded Region Growing algorithm is found to be the best since the Mean Square Error (MSE) and Standard Deviation (SD) are comparatively less. The appropriate treatment procedure obtained from an expert urologist is suggested in the GUI depending upon the size of the calculi.

Longitudinal Ultrasound Curriculum Improves Long-Term Retention Among Internal Medicine Residents.

Point-of-care ultrasound is a rapidly evolving component of internal medicine (IM) residency training. The optimal approach for teaching this skill remains unclear. We sought to determine whether the addition of a longitudinal ultrasound curriculum to a stand-alone workshop for ultrasound training improved knowledge retention in IM residents. We conducted an observational cohort study from July to December 2013. All postgraduate year (PGY)-1 IM residents attended an ultrasound workshop during orientation. Ability to identify static images of ascites, kidney, thyroid, pleural fluid, inferior vena cava, and internal jugular vein was assessed immediately after the workshop. An ultrasound curriculum, including morning report and ultrasound rounds, was initiated during the inpatient medicine rotation. PGY-1 residents were randomly assigned to participate in the longitudinal curriculum. Six months later, we conducted a follow-up survey with all PGY-1 residents. Forty-eight PGY-1 residents (67%) completed the postworkshop test and the 6-month follow-up test. Of these, 50% (24 of 48) had participated in the ultrasound curriculum. Residents not exposed to the curriculum showed a decline in the identification of ascites, pleural effusion, and internal jugular vein at 6 months (P < .05), whereas those who participated in the curriculum maintained their performance (P < .05). Six months after exposure to a longitudinal ultrasound curriculum, residents were more likely to correctly identify ultrasound images of ascites, kidney, and pleural effusion. The addition of a longitudinal ultrasound curriculum may result in improved knowledge retention in IM residents.

Global Neighbour Preserving Local Ternary Co-Occurrence Pattern (GNPLTCoP) for Ultrasound Kidney Images Retrieval

Objectives: This work proposes a feature extraction procedure named as Global Neighbour Preserving Local Ternary Co-occurrence Pattern (GNPLTCoP) in the Content Based Image Retrieval (CBIR) Task for ultrasound kidney images retrieval. Methods/Analysis: The proposed GNPLTCoP feature finds the local pattern based on the co-occurrence of first order derivatives in ternary fashion from radius 1 and radius 2 neighbourhoods of the center pixel in a small 3X3 square region. Then the pattern is classified into two in order to integrate the global information. The classification is based on the correlation between the global and local region mean intensities. Findings: The Performance of GNPLTCoP is compared with the Ternary co-occurrence Pattern (LTCoP) as it computes the co-occurrence of first order derivatives. The LTCoP considers the co-occurrence of 8 pixels in the radius 1 neighbourhood with the 8 pixel candidates in the radius 2 neighbourhood. Among the 16 pixels in the radius 2 neighbourhood, 8 pixels are used and 8 pixels information are left in the LTCoP computation. This problem is addressed in the computation of GNPLTCoP. The proposed GNPLTCoP feature is different from the LTCoP by means of considering the pixel candidates of radius 2 neighbourhood and adding in the global information. The pixel candidates of radius 2 neighbourhood are replaced by the mean value of itself and its two neighbours to preserve the neighbour pixels information. This work employs GNPLTCoP feature as a feature extraction procedure in image retrieval system consists of database of ultrasound kidney images. The performance of GNPLTCoP is compared with LTP and LTCoP. The discriminative power of GNPLTCoP is substantiated through Precision and Recall measures. Conclusion/Application: The proposed GNPLTCoP can be applied as the feature extraction procedure for other types of medical images and pattern recognition applications. Keywords: CBIR, GNPLTCoP, Local Patterns, LTCoP, Texture, Ultrasound Kidney Images

Bas-Relief Map Using Texture Analysis with Application to Live Enhancement of Ultrasound Images

For ultrasound imaging, speckle is one of the most important factors in the degradation of contrast resolution because it masks meaningful texture and has the potential to interfere with diagnosis. It is expected that researchers would explore appropriate ways to reduce the speckle noise, to find the edges of structures and enhance weak borders between different organs in ultrasound imaging. Inspired by the principle of differential interference contrast microscopy, a “bas-relief map” is proposed that depicts the texture structure of ultrasound images. Based on a bas-relief map, an adaptive bas-relief filter was developed for ultrafast despeckling. Subsequently, an edge map was introduced to enhance the edges of images in real time. The holistic bas-relief map approach has been used experimentally with synthetic phantoms and digital ultrasound B-scan images of liver, kidney and gallbladder. Based on the visual inspection and the performance metrics of the despeckled images, it was found that the bas-relief map approach is capable of effectively reducing the speckle while significantly enhancing contrast and tissue boundaries for ultrasonic images, and its speckle reduction ability is comparable to that of Kuan, Lee and Frost filters. Meanwhile, the proposed technique could preserve more intra-region details compared with the popular speckle reducing anisotropic diffusion technique and more effectively enhance edges. In addition, the adaptive bas-relief filter was much less time consuming than the Kuan, Lee and Frost filter and speckle reducing anisotropic diffusion techniques. The bas-relief map strategy is effective for speckle reduction and live enhancement of ultrasound images, and can provide a valuable tool for clinical diagnosis.

Classification of Ultrasound Kidney Images using PCA and Neural Networks

In this paper, a computer-aided system is proposed for automatic classification of Ultrasound Kidney diseases. Images of five classes: Normal, Cyst, Stone, Tumor and Failure were considered. A set of statistical features and another set of multi-scale wavelet-based features were extracted from the region of interest (ROI) of each image and the principal component analysis was performed to reduce the number of features. The selected features were utilized in the design and training of a neural network classifier. A correct classification rate of 97% has been obtained using the multi-scale wavelet-based features.