Nidhi Gupta

Bio: Nidhi Gupta is an academic researcher from Indian Institute of Information Technology, Design and Manufacturing, Jabalpur. The author has contributed to research in topics: Image segmentation & Discrete wavelet transform. The author has an hindex of 5, co-authored 10 publications receiving 139 citations. Previous affiliations of Nidhi Gupta include Indian Institutes of Information Technology.

Glioma detection on brain MRIs using texture and morphological features with ensemble learning

Abstract: The real time usage of Computer Aided Diagnosis (CAD) systems to detect brain tumors as proposed in the literature is yet to be explored. Gliomas are the most commonly found brain tumors in human. The proposed non-invasive CAD system based on brain Magnetic Resonance Imaging (MRIs) is capable of assisting radiologists and clinicians to detect not only the presence, but also the type of glioma tumors. The system is devised to work irrespective of the image pulse sequence. It uses different segmentation schemes for different pulse sequences, fusion of texture features, and ensemble classifier to perform three levels of classification. Once the tumor is detected at the first level of classification, its location is analyzed using tentorium of brain and it is classified into superatentorial or infratentorial in the next level. Based on the morphological and inherent characteristics of tumor (area, perimeter, solidity, and orientation), the system identifies tumor type at the third level of classification. The system reports average accuracy of 97.76% on JMCD (a dataset collected from local medical college) and 97.13% on BRATS datasets at the first level of classification. Average accuracy of 97.87% for astrocytomas, 94.24% for ependymoma, 96.29% for oligodendroglioma, and 98.69% for glioblastoma multiforme is observed for histologically classified JMCD dataset. The same is observed as 95.45% for low grade and 95.50% for high grade tumors in publically available BRATS dataset. The performance of the proposed CAD system is statistically examined through hypothetical Student’s t-test and Wilcoxon matched pair test. The performance of the system is also validated by domain experts for its possible real time usage.

A non-invasive and adaptive CAD system to detect brain tumor from T2-weighted MRIs using customized Otsu’s thresholding with prominent features and supervised learning

Abstract: The detection of brain tumor is a challenging task for radiologists as brain is the most complicated and complex organ. This work presents a non-invasive and adaptive method for detection of tumor from T2-weighted brain magnetic resonance (MR) images. Non-homogeneous brain MR images are enhanced by preprocessing and segmented through multilevel customization of Otsu’s thresholding technique. Several textural and shape features are extracted from the segmented image and two prominent ones are selected through entropy measure. Support vector machine (SVM) classifies MR images using prominent features. Experiments are performed on a dataset collected from MP MRI & CT Scan Centre at NSCB Medical College Jabalpur and the other from Charak Diagnostic & Research Centre Jabalpur. More than 98% accuracy is reported with 100% sensitivity for both the datasets at 99% confidence interval. The proposed system is compared with several existing methods to showcase its efficacy.

Identification of Gliomas from brain MRI through adaptive segmentation and run length of centralized patterns

Abstract: Brain tumor detection and identification of its severity is a challenging task for radiologists and clinicians. This work aims to develop a novel clinical decision support system to assist radiologists and clinicians efficiently in real-time. The proposed clinical decision support system utilizes fusion of MRI pulse sequences as each of them gives salient information for tumor identification. An adaptive thresholding is proposed for segmentation and centralized patterns are observed from LBP image of so obtained segmented image. Run length matrix extracted from these centralized patterns is used for tumor identification. The developed features successfully identify and classify tumor with Naive Bayes classifier. The proposed decision support system not only detects tumors, but also identifies its grading in terms of severity. As Glioma tumors are the most frequent among brain tumors, the proposed system is tested for the presence of low grade (Astrocytoma and Ependymoma) as well as high grade (Oligodendroglioma and Glioblastoma Multiforme) Glioma tumors on images collected from NSCB Medical College Jabalpur, India and BRATS dataset. The experiments performed on two datasets give more than 96% accuracy. The proposed decision support system is quite sensitive towards the detection and specification of tumors. All the results are verified by domain experts in real time.

Enhancement of dark images using dynamic stochastic resonance with anisotropic diffusion

Abstract: Image enhancement techniques are intended to improve the quality of an image without any kind of distortion or degradation. The literature is rich enough in this area, but there also exist some limitations. A technique is proposed for image enhancement by combining anisotropic diffusion with dynamic stochastic resonance in discrete wavelet transform domain. The method increases mean and variance of the image by the optimum iterations on low coefficients of images, which improves contrast and brightness, respectively, and simultaneously, edges also become sharper. It is well demonstrated by performing on various test images. Specifically, the adaptation and efficiency of the proposed technique for medical images are shown, because generally medical images appear contaminated with noise in terms of low illumination.

A fast and efficient computer aided diagnostic system to detect tumor from brain magnetic resonance imaging

Abstract: In this work, a simple and efficient CAD computer-aided diagnostic system is proposed for tumor detection from brain magnetic resonance imaging MRI. Poor contrast MR images are preprocessed by using morphological operations and DSR dynamic stochastic resonance technique. The appropriate segmentation of MR images plays an important role in yielding the correct detection of tumor. On examination of three views of brain MRI, it was visible that the region of interest ROI lies in the middle and its size ranges from 240 × 240 mm2 to 280 × 280 mm2. The proposed system makes effective use of this information and identifies four blocks from the desired ROI through block-based segmentation. Texture and shape features are extracted for each block of all MRIs in the training set. The range of these feature values defines the threshold to distinguish tumorous and nontumorous MRIs. Features of each block of an MRI view are checked against the threshold. For a particular feature, if a block is found tumorous in a view, then the other views are also checked for the presence of tumor. If corresponding blocks in all the views are found to be tumorous, then the MRI is classified as tumorous. This selective block processing technique improves computational efficiency of the system. The proposed technique is well adaptive and fast, and it is compared with well-known existing techniques, like k-means, fuzzy c-means, etc. The performance analysis based on accuracy and precision parameters emphasizes the effectiveness and efficiency of the proposed work.

Magnetic Resonance Imaging of the Brain and Spine

Abstract: This fifth edition continues the tradition of excellence with thorough coverage of recent trends and changes in the clinical diagnosis and treatment of CNS diseases, detailed relevant neuropathologic, genetic, and clinical findings, and how those changes relate to MRI findings. It remains a comprehensive, internationally acclaimed, state-of-the-art reference for all who have an interest in neuroradiology – trainees to experts in the field, basic science researchers, and clinicians.

Human Anatomy and Physiology I

Abstract: 3. Compare the relative diameters of the afferent and efferent arterioles and explain the significance in this size differential. The afferent arteriole is wider than the efferent arteriole which means that blood enters the glomerulus through a wider opening than the blood exiting the glomerulus, thus creating an increased “back pressure” (=hydrostatic filtration pressure). They hydrostatic pressure is higher in the glomerulus than in other capillaries. By varying the size of the afferent and/or arterioles, the glomerular filtration rate (GFR) may be increased or decreased.

Support vector machines: Theory and applications

Abstract: This chapter presents a summary of the issues discussed during the one day workshop on “Support Vector Machines (SVM) Theory and Applications” organized as part of the Advanced Course on Artificial Intelligence (ACAI ’99) in Chania, Greece [19]. The goal of the chapter is twofold: to present an overview of the background theory and current understanding of SVM, and to discuss the papers presented as well as the issues that arose during the workshop.

Brain tumor detection using statistical and machine learning method.

Abstract: Background and Objective Brain tumor occurs because of anomalous development of cells. It is one of the major reasons of death in adults around the globe. Millions of deaths can be prevented through early detection of brain tumor. Earlier brain tumor detection using Magnetic Resonance Imaging (MRI) may increase patient's survival rate. In MRI, tumor is shown more clearly that helps in the process of further treatment. This work aims to detect tumor at an early phase. Methods In this manuscript, Weiner filter with different wavelet bands is used to de-noise and enhance the input slices. Subsets of tumor pixels are found with Potential Field (PF) clustering. Furthermore, global threshold and different mathematical morphology operations are used to isolate the tumor region in Fluid Attenuated Inversion Recovery (Flair) and T2 MRI. For accurate classification, Local Binary Pattern (LBP) and Gabor Wavelet Transform (GWT) features are fused. Results The proposed approach is evaluated in terms of peak signal to noise ratio (PSNR), mean squared error (MSE) and structured similarity index (SSIM) yielding results as 76.38, 0.037 and 0.98 on T2 and 76.2, 0.039 and 0.98 on Flair respectively. The segmentation results have been evaluated based on pixels, individual features and fused features. At pixels level, the comparison of proposed approach is done with ground truth slices and also validated in terms of foreground (FG) pixels, background (BG) pixels, error region (ER) and pixel quality (Q). The approach achieved 0.93 FG and 0.98 BG precision and 0.010 ER on a local dataset. On multimodal brain tumor segmentation challenge dataset BRATS 2013, 0.93 FG and 0.99 BG precision and 0.005 ER are acquired. Similarly on BRATS 2015, 0.97 FG and 0.98 BG precision and 0.015 ER are obtained. In terms of quality, the average Q value and deviation are 0.88 and 0.017. At the fused feature based level, specificity, sensitivity, accuracy, area under the curve (AUC) and dice similarity coefficient (DSC) are 1.00, 0.92, 0.93, 0.96 and 0.96 on BRATS 2013, 0.90, 1.00, 0.97, 0.98 and 0.98 on BRATS 2015 and 0.90, 0.91, 0.90, 0.77 and 0.95 on local dataset respectively. Conclusion The presented approach outperformed as compared to existing approaches.

Deep convolutional neural networks with transfer learning for automated brain image classification

Abstract: MR brain image categorization has been an active research domain from the last decade. Several techniques have been devised in the past for MR image categorization, starting from classical to the deep learning methods like convolutional neural networks (CNNs). Classical machine learning methods need handcrafted features to perform classification. The CNNs, on the other hand, perform classification by extracting image features directly from raw images via tuning the parameters of the convolutional and pooling layer. The features extracted by CNN strongly depend on the size of the training dataset. If the training dataset is small, CNN tends to overfit after several epochs. So, deep CNNs (DCNNs) with transfer learning have evolved. The prime objective of the present work is to explore the capability of different pre-trained DCNN models with transfer learning for pathological brain image classification. Various pre-trained DCNNs, namely Alexnet, Resnet50, GoogLeNet, VGG-16, Resnet101, VGG-19, Inceptionv3, and InceptionResNetV2, were used in the present study. The last few layers of these models were replaced to accommodate new image categories for our application. These models were extensively evaluated on data from Harvard, clinical, and benchmark Figshare repository. The dataset was then partitioned in the ratio 60:40 for training and testing. The validation on the test set reveals that the pre-trained Alexnet with transfer learning exhibited the best performance in less time compared to other proposed models. The proposed method is more generic as it does not need any handcrafted features and can achieve an accuracy value of 100%, 94%, and 95.92% for three datasets. Other performance measures used in the study include sensitivity, specificity, precision, false positive rate, error, F-score, Mathew correlation coefficient, and area under the curve. The results are compared with both the traditional machine learning methods and those using CNN.