H.S. Bhadauria

Medical image denoising using adaptive fusion of curvelet transform and total variation

In medical images noise and artifacts are introduced due to the acquisition techniques and systems. Due to the noise present in the medical images, experts may not be able to draw correct and useful information from the images. The paper proposes a noise reduction method for both computed tomography (CT) and magnetic resonance imaging (MRI) which fuses the images (i) denoised by total variation (TV) method, (ii) denoised by curvelet based method and (iii) the edge information, where edge information is extracted from the noise residual of TV method by processing it through curvelet transform. The performance of the proposed method is evaluated on real brain CT and MRI images and results show significant improvement not only in noise suppression but also in edge preservation.

An integrated method for hemorrhage segmentation from brain CT Imaging

This paper presents an integrated segmentation method which combines the features of Fuzzy C-Mean (FCM) clustering and region-based active contour method. In the proposed method, FCM clustering is used to initialize the contour around the hemorrhagic region and then region-based active contour method propagates the initial contour towards the hemorrhage boundaries. Further, the FCM clustering is also used to estimate the contour propagation controlling parameters adaptively from the given image. The region-based active contour method uses the intensity information in the local regions as against the global regions in the traditional region-based active contour methods to guide the contour motion. The effectiveness of the proposed method is tested on the dataset of total 100 hemorrhagic brain CT images of 20 patients and the results are compared with region growing, FCM clustering and Chan & Vese methods. The proposed method yields the higher average values of the similarity indices namely sensitivity, specificity, accuracy and overlap metric as 79.93%, 99.10%, 84.83% and 88.84% respectively.

Classification of acute lymphoblastic leukaemia using hybrid hierarchical classifiers

In current consequence of haematology, blood cancer i.e. acute lymphoblastic leukemia is very frequently founded in medical practice, which is characterized by over activation and functional abnormality of bone marrow. The abnormality is identified through physical examination with a screening of blood smears. However, this method is error prone and labor intensive task for haematologist. Hence, haematologist needs a specific computer aided diagnostic system (CAD) that can deal with these limitations of prior systems and capable of discriminating immature leukemic cells from mature healthy cells. Thus, this work addresses the problem of segmenting a microscopic blood image into different regions, and then further analyzes those regions for localization of the immature lymphoblast cell. Further, it investigates the use of different geometrical, chromatic and statistical textures features for nucleus as well as cytoplasm and pattern recognition techniques for sub typing immature acute lymphoblasts as per FAB (French– American – British) classification. This can facilitate haematologist for acquiring essential information about prognosis and for an appropriate cure for leukemia. The exhaustive experiments have been conducted on 260 microscopic blood images (i.e. 130 normal and 130 cancerous cells) taken from ALL-IDB database. The proposed techniques consisting of the segmentation module used for segmenting the nucleus and cytoplasm of each leukocyte cell, feature extraction module, feature dimensionality reduction module that uses principal component analysis (PCA) to mapped the higher feature space to lower feature space and classification module that employs the standard classifiers, like support vector machines, smooth support vector machines, k-nearest neighbour, probabilistic neural network and adaptive neuro fuzzy inference system.

Reduction of speckle noise from medical images using principal component analysis image fusion

Images captured by different medical devices contain intrinsic artefacts, like ultrasound, CT, MRI images often contain speckle noise, which is the result of the destructive and constructive coherent summation of echoes. In these images, the speckle noise must be reduced cautiously as it also contains diagnostic information. Thus the despeckling algorithms should reduce speckle in homogeneous areas of the image and edges in the image should be preserved. In this paper a method to reduce the speckle noise is proposed which uses the concept of fusion. The performance of the proposed algorithm is quantified by calculating measures like MSE, SNR, PSNR and MSSI, which gives information about the extent of feature preservation and denoising.

An approach for leukocytes nuclei segmentation based on image fusion

Now days blood smear evaluation is a most common clinical test for the hematologists. In Most of the cases like Neutrophilia, Acute leukemia, the hematologists are eager to know about the leukocytes (WBCs) only that are detected based on the condition and blood count of the leukocytes in human body. Leukocytes cell have a complex background surrounding and its morphological structures such as nucleus and cytoplasm. Medical Imaging can help hematologist in their diagnosis by automatically extracting the nucleus from the leukocyte cell. In comparison of the existing Otsus method this paper presents the consistent approach for automatic blood cell segmentation that is high discriminated based on PCA fusion by improving Otsus with few preprocessing and post-processing techniques in order to provide an improved image for human perception and for further processing of classification of blood cells.

Analysis of effect of cycle spinning on wavelet- and curvelet-based denoising methods on brain CT images

The purpose of this paper is to carry out the assessment of effect of cycle spinning on wavelet- and curvelet-based noise reduction methods on brain CT images. In particular, multiscale curvelet- and wavelet-based denoising methods are evaluated with and without cycle spinning. This assessment is focused not only on the noise suppression but also on fine details preservation. The experimental results show that the cycle spinning-based curvelet method outperforms not only other curvelet-based methods but also the wavelet-based methods. The quality assessment parameters taken in this paper are signal-to-noise ratio (SNR), peak-signal-to-noise ratio (PSNR), universal quality index (UQI), structural similarity index metrics (SSIM), and edge keeping index (EKI).

Automatic brain MRI image segmentation using FCM and LSM

The significant objective of this paper is to produce a method that is able to delineate the object of interest or tumor region easily from the available brain MRI images. This is attained by the unification of the fuzzy c-means clustering and level set method. The method proposed performs the segmentation by smoothly exploiting the spatial function during FCM clustering. Since, we are utilizing the FCM which could prove the automaticity of the method by dividing the original image into clusters and then using one cluster for automatic initialization. This in turn helps in making the whole processing less tedious with reducing the time as well. Thereby, if considered it could be competent tool in future. Secondly, to find the contour of tumor region in the original image the proposed method uses the level set method which comes in handy in situations where the topologies of the images changes frequently by merging or splitting in two. Also, the proposed methodology makes use of variational level method in place of generic level set method which in turn eliminates one more flaw of re- initializing the contour during segmentation. When we are using the segmentation methods which are manual then it can lead to a situation where different medical experts generate different results which can also overcome by using the proposed approach.

A Survey of Noise Removal Methodologies for Lung Cancer Diagnosis

Investigation of signal and image are presently an essential step of the heart diseases processes like diagnostic, prognostic and follow-up. Lung cancer is the most intense type of cancer among every type of cancer with less rate of survival. It is exceptionally hard to examine the cancer at its initial stage. In the previous couple of years, numerous Computer aided systems have been intended to distinguish the lung cancer at its initial stage. In medical imaging, diverse types of images are being utilized yet for the detection of diagnosis of lungs. In medical imaging, detection of nodule is standout amongst the challenging tasks. Computed Tomography (CT) images are generally preferred due to less distortion, low noise and better clarity. Detecting and then curing that disease in the initial stages offers the patients with higher possibility of survival. There are different types of the noise present in the images we obtain for the lung mass detection like salt and pepper noise, Gaussian noise and speckle noise. This paper is based on quality improvement analysis of digital dental X-ray image. Removal of noise from images is the most active field of research. This paper presents the review on the lung cancer, types of noise in medical imaging and then the methods for the removal of noise.

Mammogram's denoising in spatial and frequency domain

Breast cancer is one of the most incurable diseases, which leads to the death of women globally every year. For early detection of a tumor in the breast, a basic technique called ‘Mammography’ is used, which is an x-ray analysis of breast. This work emphasizes on the proper selection of denoising techniques for the mammographic images. To achieve the objective of this work, exhaustive experiments are carried out using spatial domain filtering techniques as well as frequency domain filtering techniques on mammograms of the Mammographic Image Analysis Society (MIAS) data. The effectiveness of the techniques is evaluated in terms of Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Mean Structure Similarity Index (MSSIM), Maximum Difference (MD), Normalized Absolute Error (NAE), and Structural Content (SC). It is observed that Wavelet denoising and Median filter show better results than Adaptive Histogram Equalization (AHE), Butterworth and Frost filters.

Image super resolution based on discrete and Stationary wavelet transform using Canny edge extraction and non local mean

This paper addresses the issue of generating a high-resolution(HR) image from single low quality or low-resolution(LR) image. In this work, Discrete wavelet transform (DWT) is used with the Stationary wavelet transform (SWT) to generate or increase the resolution of the image. SWT reduces the translation invariance presence in DWT. To preserve the edges Canny edge extraction is used to get the sharper image. To interpolate the image into the intermediary stage of proposed algorithm Lanczos interpolation is used and to reduce the artifacts introduced by the DWT Non-local mean(NLM) filter has been used. The experimental result shows that the proposed algorithm gives good results based on image quality parameters as compared with the state-of-the-art works in super resolution (SR) process.