Chirag Kamal Ahuja

A novel content-based active contour model for brain tumor segmentation

Brain tumor segmentation is a crucial step in surgical and treatment planning. Intensity-based active contour models such as gradient vector flow (GVF), magneto static active contour (MAC) and fluid vector flow (FVF) have been proposed to segment homogeneous objects/tumors in medical images. In this study, extensive experiments are done to analyze the performance of intensity-based techniques for homogeneous tumors on brain magnetic resonance (MR) images. The analysis shows that the state-of-art methods fail to segment homogeneous tumors against similar background or when these tumors show partial diversity toward the background. They also have preconvergence problem in case of false edges/saddle points. However, the presence of weak edges and diffused edges (due to edema around the tumor) leads to oversegmentation by intensity-based techniques. Therefore, the proposed method content-based active contour (CBAC) uses both intensity and texture information present within the active contour to overcome above-stated problems capturing large range in an image. It also proposes a novel use of Gray-Level Co-occurrence Matrix to define texture space for tumor segmentation. The effectiveness of this method is tested on two different real data sets (55 patients - more than 600 images) containing five different types of homogeneous, heterogeneous, diffused tumors and synthetic images (non-MR benchmark images). Remarkable results are obtained in segmenting homogeneous tumors of uniform intensity, complex content heterogeneous, diffused tumors on MR images (T1-weighted, postcontrast T1-weighted and T2-weighted) and synthetic images (non-MR benchmark images of varying intensity, texture, noise content and false edges). Further, tumor volume is efficiently extracted from 2-dimensional slices and is named as 2.5-dimensional segmentation.

Hepatic arteriovenous fistulae: role of interventional radiology.

IntroductionHepatic arterial venous fistulae are abnormal communications between the hepatic artery and portal or hepatic vein and commonly occur either secondary to iatrogenic causes like liver biopsy, transhepatic biliary drainage, transhepatic cholangiogram and surgery, or following mechanical insult like blunt or penetrating trauma. Congenital fistulae are rare. Treatment is warranted as an emergency management or in the development of portal hypertension/heart failure in chronic cases. Both surgical and endovascular occlusion of the fistula can be attempted with the latter carrying low intra and post-procedure morbidity. Endovascular treatment has thus currently emerged as a minimally invasive reliable treatment option in such individuals.Methods and ResultsWe describe a short series consisting of four cases of acquired hepatic arterioportal/venous fistulae, which were referred to interventional radiology for endovascular management over the last 2 years. Three patients had arterio-portal communication and one patient had communication between the hepatic artery and middle hepatic vein. Successful embolization through the transarterial route was achieved in all four patients. A brief discussion of these cases is presented along with a relevant review of literature.ConclusionsEndovascular techniques currently form less invasive and first line treatment options in arterioportal/venous fistulae, surgery being reserved only for unsuccessful embolizations/complex fistulae.

Classification of brain tumors using PCA-ANN

The present study is conducted to assist radiologists in marking tumor boundaries and in decision making process for multiclass classification of brain tumors. Primary brain tumors and secondary brain tumors along with normal regions are segmented by Gradient Vector Flow (GVF)-a boundary based technique. GVF is a user interactive model for extracting tumor boundaries. These segmented regions of interest (ROIs) are than classified by using Principal Component Analysis-Artificial Neural Network (PCA-ANN) approach. The study is performed on diversified dataset of 856 ROIs from 428 post contrast T1- weighted MR images of 55 patients. 218 texture and intensity features are extracted from ROIs. PCA is used for reduction of dimensionality of the feature space. Six classes which include primary tumors such as Astrocytoma (AS), Glioblastoma Multiforme (GBM), child tumor-Medulloblastoma (MED) and Meningioma (MEN), secondary tumor-Metastatic (MET) along with normal regions (NR) are discriminated using ANN. Test results show that the PCA-ANN approach has enhanced the overall accuracy of ANN from 72.97 % to 95.37%. The proposed method has delivered a high accuracy for each class: AS-90.74%, GBM-88.46%, MED-85.00%, MEN-90.70%, MET-96.67%and NR-93.78%. It is observed that PCA-ANN provides better results than the existing methods.

Multiclass Brain Tumor Classification Using GA-SVM

The objective of this study is to develop a CAD system for assisting radiologists in multiclass classification of brain tumors. A new hybrid machine learning system based on the Genetic Algorithm (GA) and Support Vector Machine (SVM) for brain tumor classification is proposed. Texture and intensity features of tumors are taken as input. Genetic algorithm has been used to select the set of most informative input features. The study is performed on real 428 post contrast T1-weighted MR images of 55 patients. Primary brain tumors such as Astrocytoma (AS), Glioblastoma Multiforme (GBM), Meningioma (MEN), and child tumor-Medulloblastoma (MED) along with secondary tumor-Metastatic (MET) are classified by GA-SVM classifier. Test results showed that the GA optimization technique has enhanced the overall accuracy of SVM from 56.3 % to 91.7%. Individual class accuracies obtained are: AS-89.8%, GBM-83.3%, MEN-96%, MET-91.8%, MED-97.1%. A comparative study with earlier methods is also done. The study reveals that GA-SVM provides more accurate results than earlier methods and is tested on more diversified dataset.

A package-SFERCB-Segmentation, feature extraction, reduction and classification analysis by both SVM and ANN for brain tumors

An interactive computer aided dignostic (CAD) system for assisting inexperience young radiologists is developed. The difficulty in brain tumors classification is due to similar size, shape, location, hetrogeniety, presence of oedema, cystic and isointense regions has been the key feature of this research. Genetic Algorithm is employed as it is an easy concept and is well understood by radiologists without going in much depth of engineering.Display Omitted Brain tumors as segmented regions of interests (SROIs) by content based active contour model (CBAC).Feature extraction-intensity and texture based features.Feature reduction by Genetic Algorithm.Classification by Hybrid Models-GA-SVM and GA-ANN. The objective of this experimentation is to develop an interactive CAD system for assisting radiologists in multiclass brain tumor classification. The study is performed on a diversified dataset of 428 post contrast T1-weighted MR images of 55 patients and publically available dataset of 260 post contrast T1-weighted MR images of 10 patients. The first dataset includes primary brain tumors such as Astrocytoma (AS), Glioblastoma Multiforme (GBM), childhood tumor-Medulloblastoma (MED) and Meningioma (MEN), along with secondary tumor-Metastatic (MET). The second dataset consists of Astrocytoma (AS), Low Grade Glioma (LGL) and Meningioma (MEN). The tumor regions are marked by content based active contour (CBAC) model. The regions are than saved as segmented regions of interest (SROIs). 71 intensity and texture feature set is extracted from these SROIs. The features are specifically selected based on the pathological details of brain tumors provided by the radiologist. Genetic Algorithm (GA) selects the set of optimal features from this input set. Two hybrid machine learning models are implemented using GA with support vector machine (SVM) and artificial neural network (ANN) (GA-SVM and GA-ANN) and are tested on two different datasets. GA-SVM is proposed for finding preliminary probability in identifying tumor class and GA-ANN is used for confirmation of accuracy. Test results of the first dataset show that the GA optimization technique has enhanced the overall accuracy of SVM from 79.3% to 91.7% and of ANN from 75.6% to 94.9%. Individual class accuracies delivered by GA-SVM are: AS-89.8%, GBM-83.3%, MED-95.6%, MEN-91.8%, and MET-97.1%. Individual class accuracies delivered by GA-ANN classifier are: AS-96.6%, GBM-86.6%, MED-93.3%, MEN-96%, MET-100%. Similar results are obtained for the second dataset. The overall accuracy of SVM has increased from 80.8% to 89% and that of ANN has increased from 77.5% to 94.1%. Individual class accuracies delivered by GA-SVM are: AS-85.3%, LGL-88.8%, MEN-93%. Individual class accuracies delivered by GA-ANN classifier are: AS-92.6%, LGL-94.4%, MED-95.3%. It is observed from the experiments that GA-ANN classifier has provided better results than GA-SVM. Further, it is observed that along with providing finer results, GA-SVM provides advantage in speed whereas GA-ANN provides advantage in accuracy. The combined results from both the classifiers will benefit the radiologists in forming a better decision for classifying brain tumors.

Reversal of Low-Grade Cerebral Edema After Lactulose/Rifaximin Therapy in Patients with Cirrhosis and Minimal Hepatic Encephalopathy

OBJECTIVES:Decreased magnetization transfer ratio (MTR) in the brain characterizes cerebral edema (CE) in patients with liver cirrhosis, but the role of treatment on its reversibility has not been studied in patients who have minimal hepatic encephalopathy (MHE). This study was carried to evaluate the reversibility of CE with lactulose and rifaximin treatment in patients with MHE and role of ammonia, pro-inflammatory interleukins (IL-1, IL-6) and tumor necrosis factor (TNF)-α in its pathogenesis.METHODS:Twenty-three patients with cirrhosis (14 with MHE, 9 without MHE (NMHE)) and 6 healthy controls underwent ammonia, IL-1, IL-6, TNF-α estimation, and MTR in frontal white matter (FWM), parietal white matter (PWM), internal capsule (IC), and basal ganglia (BG).RESULTS:Ammonia was significantly higher in the cirrhosis group compared with controls and in MHE compared with the NMHE group. Ammonia correlated positively with IL-1 and IL-6. MTRs in FWM, PWM, IC, and BG were significantly lower in the MHE group compared with controls and in PWM, IC, and BG compared with the NMHE group. MHE patients showed significant MTR increase in FWM, PWM, and IC with treatment. IL-6 and ammonia had significant negative and significant positive psychometric hepatic encephalopathy score (PHES) correlation with MTR in various regions.CONCLUSIONS:This study, for the first time, demonstrated the reversibility of low-grade CE with treatment in patients with MHE. Negative correlation between ammonia, IL-6 levels, and MTR and positive correlation between PHES and MTR in MHE patients suggests the role of inflammation and ammonia in the genesis of low-grade CE.

A dual neural network ensemble approach for multiclass brain tumor classification

The present study is conducted to develop an interactive computer aided diagnosis (CAD) system for assisting radiologists in multiclass classification of brain tumors. In this paper, primary brain tumors such as astrocytoma, glioblastoma multiforme, childhood tumor-medulloblastoma, meningioma and secondary tumor-metastases along with normal regions are classified by a dual level neural network ensemble. Two hundred eighteen texture and intensity features are extracted from 856 segmented regions of interest (SROIs) and are taken as input. PCA is used for reduction of dimensionality of the feature space. The study is performed on a diversified dataset of 428 post contrast T1-weighted magnetic resonance images of 55 patients. Two sets of experiments are performed. In the first experiment, random selection is used which may allow SROIs from the same patient having similar characteristics to appear in both training and testing simultaneously. In the second experiment, not even a single SROI from the same patient is common during training and testing. In the first experiment, it is observed that the dual level neural network ensemble has enhanced the overall accuracy to 95.85% compared with 91.97% of single level artificial neural network. The proposed method delivers high accuracy for each class. The accuracy obtained for each class is: astrocytoma 96.29%, glioblastoma multiforme 96.15%, childhood tumor-medulloblastoma 90%, meningioma 93.00%, secondary tumor-metastases 96.67% and normal regions 97.41%. This study reveals that dual level neural network ensemble provides better results than the single level artificial neural network. In the second experiment, overall classification accuracy of 90.4% was achieved. The generalization ability of this approach can be tested by analyzing larger datasets. The extensive training will also further improve the performance of the proposed dual network ensemble. Quantitative results obtained from the proposed method will assist the radiologist in forming a better decision for classifying brain tumors.

Color and grey scale fusion of osseous and vascular information

Abstract Presenting an efficient form of gathering, refining and compounding the vital information fusion of osseous and vascular images together has gained increasing momentum in the past. This area has been witnessed with testing of a large variety of fusion based methods. Here in this article an underlying idea of enhancing the fusion quality and increasing the amount of information transfer from source images to fused image has been materialized. The target is achieved by applying a selected sequence of tried and validated techniques for pre- hand processing of the 2D medical data. The series of operations like denoising, enhancement, sharpening and finally the fusion of mask and DSA (Digital Subtraction Angiography) is done before they are finally fused. The results so obtained are able to present a far better visual quality than the raw data acquired from the medical institutes. With this approach of image enhancement prior to fusion we could achieve much better quality of fused images. This improved method of enhancement and fusion is able to achieve QAB/F factor as high as 0.8475 as compared to QAB/F of 0.619 achieved using the dense SIFT fusion algorithm alone by Yu Lui. The high quality of image results obtained offers a revolutionary paradigm in the diagnosis, optimization and planning of surgical or endovascular and cerebrovascular diseases. The entire work is implemented using MATLAB 2012 software

Osseous and Vascular Information Fusion using Various Spatial Domain Filters

Image fusion techniques aims at generating a composite image by integrating the complementary information from the multiple source images from the same sensor. To achieve a high amout of fidelity in the integrated image has been a major issue of concern with the researchers. It is intriguiging how simple pre-processing techniques can accelerate the fusion rate. In this article the performance evaluation of novel combination schemes of image sharpening and image fusion using spatial domain filters is presented. The performance of the proposed scheme has been verified for bone and vasscular image fusion i.e. multimodal fusion. We have reported weighted bilateral filter fusion technique as the best method in context of objective and subjective evaluation. The entire exhaustive assesment gives a QAB/F factor of 0.769.

Efficient fusion of osseous and vascular details in wavelet domain

The osseous and vascular information fusion is analyzed.Various spatial domain enhancement schemes pre-fusion results in enhanced fusion rate.Occurrence of noise and artifacts in the fused image has been tackled.High fusion rate is achieved which is in agreement with the visual results. Image fusion is important medical image processing technique which aims at fusing complimentary information between two images. A little literature has been found on pre-processing of the source images prior to fusion which directly affects the image fusion quality. In this paper we have proposed an effective image fusion method based on IHS-wavelet transform along with pre-processing of source image in a selected sequence of spatial and transform domain techniques before fusion to create a highly informative fused image. The proposed methodology has outperformed six other state-of-the-art image fusion techniques both in terms of objective evaluation and visual results on osseous and vascular 2-D data