S. Saradha

Bio: S. Saradha is an academic researcher from Periyar Maniammai University. The author has contributed to research in topics: Computer science & Artificial intelligence. The author has an hindex of 1, co-authored 1 publications receiving 300 citations.

An efficient automatic brain tumor classification using optimized hybrid deep neural network

Abstract: A significant topic of investigation in the area of medical imaging is brain tumor classification. Since precision is significant for classification, computer vision researchers have developed several approaches, but they still struggle with poor accuracy. In this paper, an automatic optimized hybrid deep neural network (OHDNN) is suggested for brain tumors. The proposed approach consists of two phases such as pre-processing and brain tumor classification. At first, the images are composed of the data, and then the collected imageries are pre-processed by using the following steps such as image enhancement and noise removal. Then the pre-processed images are fed to the classification stage. For the classification process, in this paper, OHDNN is used. The HDNN is a combination of a convolution neural network and long short-term memory (CNN-LSTM). Here, the CNN classifier is used for feature map generation and the classification process LSTM classifier is used. Besides, to improve the performance of the CNN-LSTM classifier, the parameter extant in the classifiers is randomly selected utilizing the adaptive rider optimization (ARO) algorithm. For the experimental process, an MRI image dataset is utilized. The experimental results show proposed approach attained the maximum accuracy of 97.5. • Brain tumor classification. • Convolution neural network. • Long short term memory. • Rider optimization and CNN-LSTM.

A Deep Learning Approach for Classification and Prediction of Cirrhosis Liver: Non Alcoholic Fatty Liver Disease (NAFLD)

Abstract: The early stage liver diseases prediction is an important health related research and using this kind of research easily can predict the diseases and take the remedies. The liver diseases are classified into different types such as liver cancer, liver tumor, fatty liver, hepatitis, cirrhosis etc. The early prediction of cirrhosis and earlier stages of liver failure reduce the risk. In this work proposed a new deep learning approach for prediction and classification of cirrhosis liver based on the non alcoholic fatty liver disease. The proposed work consists of different features, deep neural network and Spearman's rank correlation. The 52 features such as gray level co-occurrence matrix (GLCM) texture features, gradient co-occurrence matrix (GLGCM) texture features are used for classification and prediction. The deep neural network (DNN) used to feature prediction and classification. Based on the different features the various types of the classifications are performed. The Spearman's rank correlation used to predict the rank correlation using different layers of the DNN network. The experiment of the proposed work is performed using MRI images and datasets. The predicted dataset is evaluated using sensitivity, specificity, accuracy and precision. The predicted results are compared with existing dominated methods and it shows better results in terms of comparison parameters.

A Hybrid Approach for Prediction and Stage Wise Classification of Liver Failure

Abstract: Early stage disease prediction is an important research area in health sector and it used to helpful to give the required treatment on time. The different stages of liver failure classification are an import research to the society due to huge amount liver failure causes. The early stage of cirrhosis failure prediction reduces the risk of human life. In this research article we propose deep learning based techniques for prediction and classification using fatty liver. The new propose work is the combination of ensemble learning (EL), conventional neural network (CN) and belief neural network. So, the proposed method is called EL-CN. The EL used to predict the features and add the different features using combing all the features. The CNN is used to manage and classify the stage wise prediction and classification. The BNN is increase the accuracy and prediction rates with supporting features. The propose work EL-CN implemented using liver datasets. The liver dataset consists of MRI images and corresponding features. The propose work implemented using python programming language and used different metrics such as accuracy, specificity and sensitivity. Predicted outcomes evaluated with dominant existing works and produced better results in terms of metrics rates such as 98.8%, 98.6%, and 98.4% respectively.

A Deep Learning-based Framework for Sheep Identification System based on Facial Bio-Metrics Analysis

Abstract: Through the use of livestock, information sharing is becoming increasingly popular around the world. This study aims to see biometric face analysis be used on sheep recognition to improve sheep monitoring in the centralized database. Anchor-free region convolutional neural networks were used to detect sheep identities (AF-RCNN). Face recognition’s effectiveness as a biometric-based identification for sheep was studied utilizing reviews of face images using the deep earing approach. The method is standalone on a set of standardized facial photos from 50 sheep, using an augmentation strategy to expand the number of sheep images. The proposed method outperforms earlier methods for sheep recognition with high accuracy.

Biodiesel production through the use of different sources and characterization of oils and their esters as the substitute of diesel: A review

Abstract: The world is confronted with the twin crises of fossil fuel depletion and environmental degradation. The indiscriminate extraction and consumption of fossil fuels have led to a reduction in petroleum reserves. Petroleum based fuels are obtained from limited reserves. These finite reserves are highly concentrated in certain region of the world. Therefore, those countries not having these resources are facing a foreign exchange crisis, mainly due to the import of crude petroleum oil. Hence it is necessary to look for alternative fuels, which can be produced from materials available within the country. Although vegetative oils can be fuel for diesel engines, but their high viscosities, low volatilities and poor cold flow properties have led to the investigation of its various derivatives. Among the different possible sources, fatty acid methyl esters, known as Biodiesel fuel derived from triglycerides (vegetable oil and animal fates) by transesterification with methanol, present the promising alternative substitute to diesel fuels and have received the most attention now a day. The main advantages of using Biodiesel are its renewability, better quality exhaust gas emission, its biodegradability and the organic carbon present in it is photosynthetic in origin. It does not contribute to a rise in the level of carbon dioxide in the atmosphere and consequently to the green house effect. This paper reviews the source of production and characterization of vegetable oils and their methyl ester as the substitute of the petroleum fuel and future possibilities of Biodiesel production.

Inedible vegetable oils and their derivatives for alternative diesel fuels in CI engines: A review

Abstract: The use of inedible vegetable oils as an alternative fuel for diesel engine is accelerated by the energy crisis due to depletion of resources and increased environmental problems including the great need for edible oil as food and the reduction of biodiesel production cost, etc. Of a lot of inedible vegetable oils which can be exploited for substitute fuel as diesel fuel, seven vegetable oils, i.e., jatropha, karanja, mahua, linseed, rubber seed, cottonseed and neem oils were selected for discussion in this review paper. The application of jatropha oil as a liquid fuel for CI engine can be classified with neat jatropha oil, engine modifications such as preheating, and dual fuelling, and fuel modifications such as jatropha oil blends with other fuels, mostly with diesel fuel, biodiesel, biodiesel blends and degumming. Therefore, jatropha oil is a leading candidate for the commercialization of non-edible vegetable oils. There exists a big difference in the fuel properties of seven inedible vegetable oils and its biodiesels considered in this review. It is clear from this review that biodiesel generally causes an increase in NOx emission and a decrease in HC, CO and PM emissions compared to diesel. It was reported that a diesel engine without any modification would run successfully on a blend of 20% vegetable oil and 80% diesel fuel without damage to engine parts. This trend can be applied to the biodiesel blends even though particular biodiesel shows 40% blend. In addition, the blends of biodiesel and diesel can replace the diesel fuel up to 10% by volume for running common rail direct injection system without any durability problems.

Catalytic applications in the production of biodiesel from vegetable oils.

Abstract: The predicted shortage of fossil fuels and related environmental concerns have recently attracted significant attention to scientific and technological issues concerning the conversion of biomass into fuels. First-generation biodiesel, obtained from vegetable oils and animal fats by transesterification, relies on commercial technology and rich scientific background, though continuous progress in this field offers opportunities for improvement. This review focuses on new catalytic systems for the transesterification of oils to the corresponding ethyl/methyl esters of fatty acids. It also addresses some innovative/emerging technologies for the production of biodiesel, such as the catalytic hydrocracking of vegetable oils to hydrocarbons. The special role of the catalyst as a key to efficient technology is outlined, together with the other important factors that affect the yield and quality of the product, including feedstock-related properties and various system conditions.