Journal Articles
Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/19884
Browse
10 results
Search Results
Item Image Analysis of Nuclei Histopathology Using Deep Learning: A Review of Segmentation, Detection, and Classification(Springer, 2023) Kadaskar, M.; Patil, N.Deep learning has recently advanced in its applicability to computer vision challenges, and medical imaging has become the most used technique in histopathology image analysis. Nuclei instance segmentation, detection, and classification are one such task. Reliable analysis of these image slides is critical in cancer identification, treatment, and care. Researchers have recently been interested in this issue. This study reviews the categorization and investigation of strategies utilized in recent works to improve the effectiveness of automated nuclei segmentation, detection, and classification in histopathology images. It critically examines state-of-the-art deep learning techniques, analyzes the trends, identifies the challenges, and highlights and helps with the future directions for research. The taxonomy includes deep learning techniques, enhancement, and optimization methods. The survey findings will help to overcome the challenges of nuclei segmentation, detection, and classification while improving the performance of models and, thus, aid future research plans. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.Item Machine learning techniques for periodontitis and dental caries detection: A narrative review(Elsevier Ireland Ltd, 2023) Radha, R.C.; Raghavendra, B.S.; Subhash, B.V.; Rajan, J.; Narasimhadhan, A.V.Objectives: In recent years, periodontitis, and dental caries have become common in humans and need to be diagnosed in the early stage to prevent severe complications and tooth loss. These dental issues are diagnosed by visual inspection, measuring pocket probing depth, and radiographs findings from experienced dentists. Though a glut of machine learning (ML) algorithms has been proposed for the automated detection of periodontitis, and dental caries, determining which ML techniques are suitable for clinical practice remains under debate. This review aims to identify the research challenges by analyzing the limitations of current methods and how to address these to obtain robust systems suitable for clinical use or point-of-care testing. Methods: An extensive search of the literature published from 2015 to 2022 written in English, related to the subject of study was sought by searching the electronic databases: PubMed, Institute of Electrical and Electronics Engineers (IEEE) Xplore, and ScienceDirect. Results: The initial electronic search yielded 1743 titles, and 55 studies were eventually included based on the selection criteria adopted in this review. Studies selected were on ML applications for the automatic detection of periodontitis and dental caries and related dental issues: Apical lessons, Periodontal bone loss, and Vertical root fracture. Conclusion: While most of the ML-based studies use radiograph images for the detection of periodontitis and dental caries, few pieces of the literature revealed that good diagnostic accuracy could be achieved by training the ML model even with mobile photos representing the images of dental issues. Nowadays smartphones are used in every sector for different applications. Training the ML model with as many images of dental issues captured by the smartphone can achieve good accuracy, reduce the cost of clinical diagnosis, and provide user interaction. © 2023 Elsevier B.V.Item Machine learning-based approaches to enhance the soil fertility—A review(Elsevier Ltd, 2024) Sujatha, M.; Jaidhar, C.D.Agriculture plays an imperative role in many countries’ economies and is a substantive source of survival. The variation in a soil nutrient decreases crop yield. An accurate soil fertility classification and application of fertilizers are essential for enhancing crop productivity. Currently, soil fertility levels are assessed through laboratory testing of soil samples, and fertilizers are applied randomly. This traditional practice increases fertilization costs and causes environmental pollution. Thus, it is necessary to develop robust and inexpensive soil fertility classification and fertilizer application. This study identifies the machine learning (ML) or deep learning-based soil fertility classifications. A comprehensive review is conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The purpose of this study is to examine different approaches that researchers use to predict or classify soil fertility. It also discusses the fertilizer recommendation developed by the researchers. The earlier research showed that ML-based approaches could accurately classify soil fertility. Furthermore, this study discusses the importance of soil nutrients and preventive measures to be taken on the imbalance of soil nutrients. This study explores research gaps and challenges in soil fertility classification and fertilizer recommendation systems. Most studies predicted the fertility levels of soil parameters, whereas a few researchers classified soil fertility. Few researchers recommended fertilizers for soil nutrient depletion. Most studies relied on expensive laboratory measurements or regional soil data collected from satellites. Based on the identified research gaps, this study suggests potential future research possibilities in soil fertility classification and the recommendation of fertilizers. It aims to develop a low-cost soil fertility classifier to prescribe fertilizers. The developed model can help farmers to enhance soil fertility with reduced fertilization costs. © 2023 Elsevier LtdItem An intelligent content-based image retrieval system for clinical decision support in brain tumor diagnosis(Springer London, 2013) Arakeri, M.P.; Guddeti, G.Accurate diagnosis is crucial for successful treatment of the brain tumor. Accordingly in this paper, we propose an intelligent content-based image retrieval (CBIR) system which retrieves similar pathology bearing magnetic resonance (MR) images of the brain from a medical database to assist the radiologist in the diagnosis of the brain tumor. A single feature vector will not perform well for finding similar images in the medical domain as images within the same disease class differ by severity, density and other such factors. To handle this problem, the proposed CBIR system uses a two-step approach to retrieve similar MR images. The first step classifies the query image as benign or malignant using the features that discriminate the classes. The second step then retrieves the most similar images within the predicted class using the features that distinguish the subclasses. In order to provide faster image retrieval, we propose an indexing method called clustering with principal component analysis (PCA) and KD-tree which groups subclass features into clusters using modified K-means clustering and separately reduces the dimensionality of each cluster using PCA. The reduced feature set is then indexed using a KD-tree. The proposed CBIR system is also made robust against misalignment that occurs during MR image acquisition. Experiments were carried out on a database consisting of 820 MR images of the brain tumor. The experimental results demonstrate the effectiveness of the proposed system and show the viability of clinical application. © 2013, Springer-Verlag London.Item A fast and novel approach based on grouping and weighted mRMR for feature selection and classification of protein sequence data(Inderscience Publishers, 2020) Kaur, K.; Patil, N.The analysis of protein sequences under bioinformatics has gained wide importance in research area. Newly added protein sequences can be analysed using existing proteins and converting them into feature vector form. However, it emerges as a challenging task to deal with huge number of features obtained using sequence encoding techniques. Since all the features obtained are not actually required, a three-stage feature selection approach has been proposed. In the first stage, features are ranked and most irrelevant features are removed; in the second stage, conflicting features are grouped together; and in third stage, a fast approach based on weighted Minimum Redundancy Maximum Relevance (wMRMR) has been proposed and applied on grouped features. Different classification methods are used to analyse the performance of the proposed approach. It is observed that the proposed approach has increased classification accuracy results and reduced time consumption in comparison to the state-of-the-art methods. © 2020 Inderscience Enterprises Ltd.Item Intraday Stock Prediction Based on Deep Neural Network(Springer, 2020) Naik, N.; Mohan, B.R.Predicting stock price movements is difficult due to the speculative nature of the stock market.Accurate predictions of stock prices allow traders to increase their profits. Stock prices react when receiving new information.During the trading day, it is difficult to understand the up and down movements signaled by stock prices. This paper addresses the problem of fluctuations in stock prices. We proposed the method to identify stock movement trend in data, and this method considered the combination of candlestick data and technical indicator values. The outcome of this method is given as inputs to a deep neural network (DNN) to classify a stock price’s up and down movements. National Stock Exchange, India, datasets are considered for an experiment from the years 2008 to 2018. The work is carried out using H2O deep learning on an RStudio platform. Experimental results are compared with a three-layer artificial neural network (ANN) model. The proposed five-layer DNN model outperforms state-of-the-art methods by 8–11% in predicting up and down movements of a given stock. © 2019, The National Academy of Sciences, India.Item A machine-learning approach for classifying Indian internet shoppers(Henry Stewart Publications, 2022) Majhi, R.; Sugasi, R.P.This paper identifies the key factors that influence Indian consumers to shop online. The study uses data collected via questionnaire survey to segment consumers with shared behaviours into groups, with the results of this clustering used to train radial basis function neural networks, decision trees and random forest models. The performance of these classification models is then assessed and compared with the conventional statistical-based naïve Bayes method and logistic regression. The study finds that the random forest method provides the greatest accuracy for the segmentation of online consumers, followed by naïve Bayes and decision trees methods. The behavioural patterns identified in this study may be leveraged in real-world situations. © 2022, Henry Stewart Publications. All rights reserved.Item Utilizing Deep Learning Models and Transfer Learning for COVID-19 Detection from X-Ray Images(Springer, 2023) Agrawal, S.; Venkatesh, V.; Nara, M.; Patil, N.COVID-19 has been a global pandemic. Flattening the curve requires intensive testing, and the world has been facing a shortage of testing equipment and medical personnel with expertise. There is a need to automate and aid the detection process. Several diagnostic tools are currently being used for COVID-19, including X-Rays and CT-scans. This study focuses on detecting COVID-19 from X-Rays. We pursue two types of problems: binary classification (COVID-19 and No COVID-19) and multi-class classification (COVID-19, No COVID-19 and Pneumonia). We examine and evaluate several classic models, namely VGG19, ResNet50, MobileNetV2, InceptionV3, Xception, DenseNet121, and specialized models such as DarkCOVIDNet and COVID-Net and prove that ResNet50 models perform best. We also propose a simple modification to the ResNet50 model, which gives a binary classification accuracy of 99.20% and a multi-class classification accuracy of 86.13%, hence cementing the ResNet50’s abilities for COVID-19 detection and ability to differentiate pneumonia and COVID-19. The proposed model’s explanations were interpreted via LIME which provides contours, and Grad-CAM, which provides heat-maps over the area(s) of interest of the classifier, i.e., COVID-19 concentrated regions in the lungs, and realize that LIME explains the results better. These explanations support our model’s ability to generalize. The proposed model is intended to be deployed for free use. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.Item ANet: Nuclei Instance Segmentation and Classification with Attention-Based Network(Springer, 2024) Kadaskar, M.; Patil, N.The segmentation and classification of nuclei in haematoxylin and eosin-stained images is critical for diagnosing cancer and other disorders. Developing automated methods is necessary for the quantitative analysis of whole-slide images and further downstream analysis. However, many challenges are to be solved, such as varying morphology and observer differences. To address these concerns, we present ANet, an encoder–decoder structure based on attention mechanisms for nuclear segmentation and classification that makes use of information in high-dimensional features improved by attention. These blocks generate meaningful feature activation and eliminate irrelevant information to produce finer maps. It segments the touching, clustered, and overlapping nuclei and classifies them using upsampling branches. Our method includes components such as PreAct-ResNet50, residual attention, convolutional block attention module, and dense attention unit. We demonstrate how our approach achieves cutting-edge performance on several multi-tissue histopathology datasets such as Kumar, CoNSeP, and CPM17. We also demonstrate our model’s generalization capabilities on other combinations of datasets, including CPM15 and TNBC. ANet demonstrates a notable improvement of 1.14%, 2.70%, 1.41%, and 1.29% in Dice, AJI, SQ, and PQ scores, respectively, for the CPM17 dataset. In addition, it achieves a 1.18% improvement in AJI score for the Kumar dataset. Despite the inherent challenges in nuclei classification within the CoNSeP dataset, ANet yields outstanding results, showcasing a substantial improvement of 9.74%, 3.97%, and 0.80% in F1 scores for the inflammatory, spindle, and miscellaneous classes. Furthermore, ANet exhibits strong generalization across the CPM dataset, TNBC, and Combined CoNSeP, with improvements observed in the majority of metrics. The given improvement is justifiable, as are the interpretable visual results. The proposed method is of great potential for analyzing histopathology images, demonstrated by an increment of performance in multiple metrics. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2024.Item ADConv-Net: Advanced Deep Convolution Neural Network for COVID-19 Diagnostics Using Chest X-Ray and CT Images(Springer, 2025) Kumar, S.; Bhowmik, B.The worldwide COVID-19 epidemic has emerged as a significant concern, affecting daily lives and underscoring the importance of early diagnosis for effective treatment in medical and healthcare settings. Current diagnostic testing for COVID-19 is sluggish, typically requiring hours to get results. Detection of COVID-19 from medical imaging presents a challenging task that has gained substantial interest from experts worldwide. Essential imaging modalities for diagnosing COVID-19 include chest X-rays and computed tomography (CT) scans. By contrast, most of the chest radiography can be completed in within fifteen minutes. Thus, employing chest radiography gives a possibility for early and reliable diagnosis of COVID-19, intending to relieve therapeutic obstacles for patients and speed up the diagnostic process. Recently, deep learning (DL) techniques have been shown to be effective in image-based diagnostics. This paper proposed an advanced deep convolution neural network (ADConv-Net) for COVID-19 detection and categorization using chest X-ray and CT images. The proposed technique is not only capable of recognizing critical connections and similarities in image classification, but also leads to improved diagnostic accuracy. The proposed model undergoes thorough evaluation for standard performance metrics. After evaluation, the ADConv-Net model achieves high accuracies of 98.84% and 97.25% in training and testing for X-ray images and 99.41% and 98.87% in training and testing for CT images, respectively. Additionally, the proposed model demonstrates strong performance, with AUC values of 0.993 and 0.996 for X-ray and CT images, respectively. Further, the model also introduces a heatmap approach for displaying COVID-19 disease areas. Subsequently, radiologists can find COVID-19 disorders in chest X-ray and CT images with this approach. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.