Search

Brain tumors are potentially fatal presence of cancer cells over a human brain, and they need to be segmented for accurate and reliable planning of diagnosis. Segmentation process must be carried out in different regions based on which the stages of cancer can be accurately derived. Glioma patients exhibit a different level of challenge in terms of cancer or tumors detection as the Magnetic Resonance Imaging (MRI) images possess varying sizes, shapes, positions, and modalities. The scanner used for sensing the location of tumors cells will be subjected to additional protocols and measures for accuracy, in turn, increasing the time and affecting the performance of the entire model. In this view, Convolutional Neural Networks deliver suitable models for efficient segmentation and thus delivered promising results. The previous strategies and models failed to adhere to diversity of sizes and shapes, proving to be a well-established solution for detecting tumors of bigger size. Tumors tend to be smaller in size and shape during their premature stages and they can easily evade the algorithms of Convolutional Neural Network (CNN). This proposal intends to furnish a detailed model for sensing early stages of cancer and hence perform segmentation irrespective of the current size and shape of tumors. The size of networks and layers will lead to a significant weightage when multiple kernel sizes are involved, especially in multiresolution environments. On the other hand, the proposed model is designed with a novel approach including a dilated convolution and level-based learning strategy. When the convolution process is dilated, the process of feature extraction deals with multiscale objective and level-based learning eliminates the shortcoming of previous models, thereby enhancing the quality of smaller tumors cells and shapes. The level-based learning approach also encapsulates the feature reconstruction processes which highlights the sensing of small-scale tumors growth. Inclusively, segmenting the images is performed with better accuracy and hence detection becomes better when compared to that of hierarchical approaches. [ABSTRACT FROM AUTHOR]

Due to their location, malignant brain tumors are one of humanity's greatest killers, among these tumors, gliomas are the most common. The early detection of gliomas can contribute to the design of proper treatment schemes and, thus, improve the survival rate of patients. However, it is a challenging task to detect the gliomas within the complex structure of the brain. The conventional artificial diagnosis is time-consuming and relies on the clinical experience of radiologists. To detect gliomas more efficiently, this paper proposes a noninvasive automatic diagnosis system for gliomas based on the machine learning methods. First, image standardization, including size normalization and background removal, is applied to produce standard images; then, the modified dynamic histogram equalization is implemented to enhance the low-contrast standard brain images, and skull removal based on outlier detection is presented. Furthermore, hybrid features, including gray-level co-occurrence matrix, pyramid histogram of the oriented gradient, modified completed local binary pattern, and intensity-based features are extracted together from the enhanced images, and their dimensions are reduced by principal component analysis. Kernel support vector machine (KSVM) combined with the particle swarm optimization is eventually adopted to train classifiers; in this paper, brain magnetic resonance imaging images are labeled with normal, glioma, and other. The experimental results show that the accuracy, sensitivity, and specificity of the proposed method can reach 98.36%, 99.17%, and 97.83%, respectively, which indicates that the proposed method performs better than many current systems.

Tumor grown in the human brains is one of the significant reasons that lead to loss of lives globally. Tumor is malignant collection of cells that grow in the human body. If these tumors grow in the brain, then they are called as brain tumors. Every year large number of human lives are lost due to this disease. Early detection of the disease might save the lives but requires experienced clinicians and diagnostic procedure that requires time and is very expensive. Therefore, there is a requirement for a robust system that automates the process of tumor identification. The idea behind this paper is to diagnose brain tumors by identifying the affected regions from the brain MRI images using machine learning approaches. In the proposed approach, prominent features of the tumor images are collected by passing them through a pre-trained Convolutional Network, VGG16. We observe that SVM gives better accuracy than other models. Though we achieve 84% accuracy, we feel the performance is not satisfactory. To make the model more robust, we obtain the most discriminant features, by applying Linear Discriminant Analysis (LDA) on the features obtained from VGG16. We use different conventional models like logistic regression, K-Nearest neighbor classifier (KNN), Perceptron learning, Multi Layered Perceptron (MLP) and Support Vector Machine (SVM) for the comparison study of the tumor image classification task. The proposed model leads to an accuracy of 100% as deep features extract important characteristics of the data and further LDA projects the data onto the most discriminant directions. [ABSTRACT FROM AUTHOR]