Manaswini Pradhan

A collection of genetic disorders and various types of abnormalities in the metabolism lead to cancer, a fatal disease. Lung and colon cancer are found to be main causes of death and infirmity in people. When choosing the best course of treatment, the diagnosis of these tumors is usually the most important consideration. This study's main objectives are to classify lung cancer and its severity, as well as to recognize malignant lung nodules. The suggested approach additionally classifies the stages of lung cancer in order to recognize lung nodules. The convolutional neural network (CNN) is used to detect lung nodules, identifying a nodule which is accurately segmented and classified. The suggested method is separated into dual parts: primarily, it classifies normal and abnormal behavior, and the subsequent one classifies the different stages of lung cancer. Texture and intensity-based features are extracted during the classification stage. When compared to other methods such as nested long short-term memory (LSTM)+ CNN, the hybrid CNN LSTM obtains superior outcomes in terms of accuracy (99.35%), specificity (99.30%), sensitivity (99.32%), and F1-score (99.29%).

A novel automated method for segmenting retinal layers in three‐dimensional (3D) space from spectral domain optical coherence tomography (SD‐OCT) images. Compared to 2D segmentation, 3D segmentation uses more data and produces findings that are more accurate and reliable. The class‐specific area of interest (ROI) choice and three important reference class approximations make the suggested technique precise, effective, and reliable. In the first step, contours are detected based on gradient intensity. To choose a smaller region of interest (ROI), the second stage entails acquiring the identified boundary neighbour B scan data for the selected ROI by categorising the problem as a graph problem. The third stage involves locating edge pixels using Canny Edge Detection from nodes. In order to calculate the edge weight of a histogram, slope similarity to the reference line and node characteristics are considered. The fourth phase boundary is precisely found by Dijkstra's shortest path algorithm. The accuracy of the method was tested based on 288 B scans of 12 patients (ten normal macular degeneration (AMD) subjects and 2 age‐related subjects from two different institutions). Five recent automated procedures are compared with the results to further validate the findings of the fifth phase. The outcomes demonstrate a mean original mean square error (RMSE) for each of the cut‐off values, which are 2.82, 4.88, 2.03, 3.77, and 0.64 pixels, respectively. As can be seen, the suggested strategy outperforms the existing models' significantly with a return on investment of 0.26 pixels.

This study mainly focuses on pre-processing the HAM10000 and BCN20000 skin lesion datasets to select important features that will drive for proper skin cancer classification. In this work, three feature fusion strategies have been proposed by utilizing three pre-trained Convolutional Neural Network (CNN) models, namely VGG16, EfficientNet B0, and ResNet50 to select the important features based on the weights of the features and are coined as Adaptive Weighted Feature Set (AWFS). Then, two other strategies, Model-based Optimized Weighted Feature Set (MOWFS) and Feature-based Optimized Weighted Feature Set (FOWFS), are proposed by optimally and adaptively choosing the weights using a meta-heuristic artificial jellyfish (AJS) algorithm. The MOWFS-AJS is a model-specific approach whereas the FOWFS-AJS is a feature-specific approach for optimizing the weights chosen for obtaining optimal feature sets. The performances of those three proposed feature selection strategies are evaluated using Decision Tree (DT), Naïve Bayesian (NB), Multi-Layer Perceptron (MLP), and Support Vector Machine (SVM) classifiers and the performance are measured through accuracy, precision, sensitivity, and F1-score. Additionally, the area under the receiver operating characteristics curves (AUC-ROC) is plotted and it is observed that FOWFS-AJS shows the best accuracy performance based on the SVM with 94.05% and 94.90%, respectively, for HAM 10000 and BCN 20000 datasets. Finally, the experimental results are also analyzed using a non-parametric Friedman statistical test and the computational times are recorded; the results show that, out of those three proposed feature selection strategies, the FOWFS-AJS performs very well because its quick converging nature is inculcated with the help of AJS.

In this chapter, the authors propose an adaptive bacterial foraging optimization (ABFO) algorithm for selection of features to increase the classification accuracy of heart disease diagnosis. In this approach, noises contained in the cardiac image are removed using median filter initially. Then, GLCM features are extracted from the cardiac image. Among the extracted features, optimal features are chosen using the ABFO algorithm. These selected features are then input to the classifier, which is a support vector neural network (RBNN). The classifier classifies the image into normal and abnormal. Simulation results show that the ABFO-based RBNN performs better than the conventional RBNN, ANN, KNN, and SVM in terms of accuracy.

In this chapter, chimp optimization algorithm (ChOA) a bio-inspired optimized technique are proposed for selection of features to increase the classification accuracy of heart disease diagnosis. In this approach, noises contained in the cardiac image are removed using median filter initially. Then, GLCM features are extracted from the cardiac image. Among the extracted features, optimal features are chosen using ChOA algorithm. These selected features are taken as input to the classifier. In this approach, support vector neural network (SVNN) is used as a classifier. The classifier classifies the image into normal and abnormal. Simulation results depict that the ChOA-based SVNN performs better than the conventional SVNN, ANN, KNN, and SVM in terms of accuracy.

The histopathology images are effective in identifying the location and level of cancer. In this chapter, a novel model is implemented for an automatic classification of histopathological images related to lung tissues. Initially, the color normalization technique is applied for improving the contrast of the histopathological images, which are acquired from the LC25000 lung histopathological image dataset. Additionally, the cancer segmentation is accomplished utilizing saliency driven region edge-based top-down level set (SDREL). Further, the feature descriptors—Alexnet and Gray Level Co-Occurrence Matrix (GLCM) features—were used for extracting the feature vectors from the segmented histopathology images. Lastly, the enhanced grasshopper optimization algorithm (EGOA) and random forest classifier were used for optimal feature selection and lung tissue classification. The simulation result shows that the EGOA-random forest model obtained 98.50% of accuracy on the LC25000 lung histopathological image dataset.

Determining lung tumor level and reducing patient mortality is a challenging task. So, the identification of benign or malignant lung nodules requires efficient and accurate methods of lung nodule diagnosis. For achieving this aim, in this paper, an adaptive radial basis neural network (RBNN) is proposed. Initially, the texture features are extracted and the extracted features are fed to the classifier to classify a nodule as benign or malignant nodule. In addition, the radial basis neural network is enhanced by using red deer optimization algorithm, which is used for optimal parameter selection. The effectiveness of the proposed approach is calculated by using different evaluation metrics. The effectiveness of the classification performance is compared with existing algorithms.

IndianJournals.com - Gateway to access, disperse and preserve knowledge!