Robinson Thamburaj

Automated Segmentation of Nuclei in Breast Cancer Histopathology Images.

The process of Nuclei detection in high-grade breast cancer images is quite challenging in the case of image processing techniques due to certain heterogeneous characteristics of cancer nuclei such as enlarged and irregularly shaped nuclei, highly coarse chromatin marginalized to the nuclei periphery and visible nucleoli. Recent reviews state that existing techniques show appreciable segmentation accuracy on breast histopathology images whose nuclei are dispersed and regular in texture and shape; however, typical cancer nuclei are often clustered and have irregular texture and shape properties. This paper proposes a novel segmentation algorithm for detecting individual nuclei from Hematoxylin and Eosin (H&E) stained breast histopathology images. This detection framework estimates a nuclei saliency map using tensor voting followed by boundary extraction of the nuclei on the saliency map using a Loopy Back Propagation (LBP) algorithm on a Markov Random Field (MRF). The method was tested on both whole-slide images and frames of breast cancer histopathology images. Experimental results demonstrate high segmentation performance with efficient precision, recall and dice-coefficient rates, upon testing high-grade breast cancer images containing several thousand nuclei. In addition to the optimal performance on the highly complex images presented in this paper, this method also gave appreciable results in comparison with two recently published methods—Wienert et al. (2012) and Veta et al. (2013), which were tested using their own datasets.

Detection of Plasmodium Falciparum in Peripheral Blood Smear Images

Malaria is a mosquito-borne infectious disease caused by the parasite Plasmodium, which requires accurate and early diagnosis for effective containment. In order to diagnose malaria in a patient, timely detection of malaria parasites in blood smear images is vital. The traditional methods are time–consuming, tedious and the quality of detection is highly subjective to the individual who performs the analysis. These results can clearly be improved upon by using image processing techniques. The malaria parasite appears in four stages, namely the ring, trophozoite, schizont, and gametocyte. The ring and the gametocyte stage are the ones seen in a peripheral blood smear and hence detecting these two stages, would help in the accurate diagnosis of malaria. The proposed work aims at automating the analysis of the blood smear images using appropriate segmentation techniques, thereby detecting infected red blood cells as well as the gametocytes found in the blood.

Automatic Detection of Tubules in Breast Histopathological Images

Histopathological examination of tissues enables pathologists to quantify the morphological features and spatial relationships of the tissue components. This process aids them in detecting and grading diseases, such as cancer. Quite often this system leads to observer variability and therefore affects patient prognosis. Hence quantitative image-analysis techniques can be used in processing the histopathology images and to perform automatic detection and grading. This paper proposes a segmentation algorithm to segment all the objects in a breast histopathology image and identify the tubules in them. The objects including the tubules and fatty regions are identified using K-means clustering. Lumen belonging to tubules is differentiated from the fatty regions by detecting the single layered nuclei surrounding them. This is done through grid analysis and level set segmentation. Identification of tubules is important because the percentage of tubular formation is one of the parameters used in breast cancer detection and grading.

Segmentation of sputum smear images for detection of tuberculosis bacilli

Background Tuberculosis (TB) is a common and lethal infectious disease, which requires accurate and early diagnosis for effective containment. Essential for the diagnosis of pulmonary infection is the detection of the bacilli through the manual microscopic examination of ZN-stained sputum smear, which is a time-consuming, complex process necessitating at least 8-10 minutes per slide. Moreover, the quality of the detection is highly subjective to the individual who performs the analysis. These results can clearly be improved upon by using image processing techniques. The proposed work uses the segmentation techniques to automate the analysis of the sputum smear images and to detect the presence of tuberculosis bacilli in them.

Automated Nuclear Pleomorphism Scoring in Breast Cancer Histopathology Images Using Deep Neural Networks

Scoring the size/shape variations of cancer nuclei nuclear pleomorphism in breast cancer histopathology images is a critical prog-nostic marker in breast cancer grading and has been subject to a con-siderable amount of observer variability and subjectivity issues. In spite of a decade long histopathology image analysis research, automated as-sessment of nuclear pleomorphism remains challenging due to the com-plex visual appearance and huge variability of cancer nuclei.This study proposes a practical application of the deep belief based deep neural net-work DBN-DNN model to determine the nuclear pleomorphism score of breast cancer tissue. The DBN-DNN network is trained to classify a breast cancer histology image into one of the three groups: score 1, score 2 and score 3 nuclear pleomorphism by learning the mean and standard deviation of morphological and texture features of the entire nuclei population contained in a breast histology image. The model was trained for features from automatically-segmented nuclei from 80 breast cancer histopathology images selected from publicly available MITOS-ATYPIA dataset. The classification accuracy of the model on the training and testing datasets was found to be 96i¾?% and 90i¾?% respectively.

Detection of Overlapping Tuberculosis Bacilli in Sputum Smear Images

Tuberculosis (TB) is a common and lethal infectious disease caused by a germ (bacterium) called Mycobacterium tuberculosis. Early diagnosis of the disease is one of the primary challenges in curtailing its spread and is a critical step in the TB-Control Program worldwide. Among the most common methods used in the diagnosis of TB is the manual microscopic examination of a ZN-stained sputum smear which is a time-consuming and error-prone process. The diagnosis crucially depends on the number of viable or dormant mycobacteria in the sputum, which are seen as red colored rod-shaped objects in the smear under a microscope. This also means that the mycobacteria have to be detected accurately in order to arrive at the correct count, the accuracy of which could be affected when there are overlapping bacilli in the images. The use of Image Analysis in the detection of the mycobacteria will introduce a paradigm shift. The proposed work identifies such overlapping mycobacteria and uses techniques to total them accurately, which is an extension of an earlier work focusing only on segmentation of the tiny organisms. Normal bacilli are just 2-4 micrometers in length and 0.2-0.5 um in width. All the organisms that fall above their average size or show a variation in the ratio of the major-to-minor axis are identified to be overlapping mycobacteria, which are then used for further analysis. The count of mycobacteria that overlap is computed by obtaining the branch points in the skeleton of the overlapping object. The dataset used in the research consisted of eighty images, which were tested using a prototype application that achieved a success rate of 70%.

Accepting H-Array Splicing Systems

In [3], Tom Head defined splicing systems motivated by the behaviour of DNA sequences. The splicing system makes use of a new operation, called splicing on strings of symbols. Paun et al. [7] extended the definition of Head and defined extended H systems which are computationally universal.

Parallel Contextual Array Insertion Deletion P System

We introduce a new P system model called as parallel contextual array insertion deletion P system, based on the modified row and column contextual rules of parallel contextual array grammar. We can generate a family of two-dimensional picture languages using this P system. We discuss some properties of this P system and find its generating power by comparing this new family of languages with that of certain other well known families of two-dimensional picture languages.

Parallel Contextual Hexagonal Array P Systems

We introduce new P system models, called as external and internal parallel contextual hexagonal array P systems, based on the external and internal parallel contextual hexagonal array grammars. We can generate hexagonal arrays using these P system models with the help of Z-direction, X-direction and Y-direction external or internal parallel contextual hexagonal array rules. We discuss some basic properties of these P systems and give some comparison results in terms of their generative powers.

On Accepting Hexagonal Array Splicing System with Respect to Directions

A new approach to hexagonal array splicing system with respect to directions in uniform and nonuniform ways as an accepting device was proposed in Sheena Christy and Thomas (Proceedings of national conference on mathematics and computer applications 2015). In this paper, we prove the following results (i) \( \mathcal{L}({\text{AHexASSD}}) - {\text{HKAL}} \ne \phi \) (ii) \( \mathcal{L}({\text{AHexASSD}}) - {\text{CT0LHAL}} \ne \phi \) (iii) \( \mathcal{L}_{n} ({\text{AHexASSD}}) \subset \mathcal{L}({\text{AHexASSD}}) \)