Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Image classification is a complex process that may be affected by many factors. This paper examines current practices, problems, and prospects of image classification. The emphasis is placed on the summarization of major advanced classification approaches and the techniques used for improving classification accuracy. In addition, some important issues affecting classification performance are discussed. This literature review suggests that designing a suitable image-processing procedure is a prerequisite for a successful classification of remotely sensed data into a thematic map. Effective use of multiple features of remotely sensed data and the selection of a suitable classification method are especially significant for improving classification accuracy. Non-parametric classifiers such as neural network, decision tree classifier, and knowledge-based classification have increasingly become important approaches for multisource data classification. Integration of remote sensing, geographical information systems (GIS), and expert system emerges as a new research frontier. More research, however, is needed to identify and reduce uncertainties in the image-processing chain to improve classification accuracy.

/pdf/a-survey-of-image-classification-methods-and-techniques-for-5bc5ltr9dq.pdf

A survey of image classification methods and techniques for improving classification performance

A wide range of methods for analysis of airborne- and satellite-derived imagery continues to be proposed and assessed. In this paper, we review remote sensing implementations of support vector machines (SVMs), a promising machine learning methodology. This review is timely due to the exponentially increasing number of works published in recent years. SVMs are particularly appealing in the remote sensing field due to their ability to generalize well even with limited training samples, a common limitation for remote sensing applications. However, they also suffer from parameter assignment issues that can significantly affect obtained results. A summary of empirical results is provided for various applications of over one hundred published works (as of April, 2010). It is our hope that this survey will provide guidelines for future applications of SVMs and possible areas of algorithm enhancement.

https://www.aboutgis.com/Publications/Mountrakis_SVM_review_in_remote_sensing_ISPRS2010.pdf

Support vector machines in remote sensing: A review

Gliomas are the most common primary brain malignancies, with different degrees of aggressiveness, variable prognosis and various heterogeneous histologic sub-regions, i.e., peritumoral edematous/invaded tissue, necrotic core, active and non-enhancing core. This intrinsic heterogeneity is also portrayed in their radio-phenotype, as their sub-regions are depicted by varying intensity profiles disseminated across multi-parametric magnetic resonance imaging (mpMRI) scans, reflecting varying biological properties. Their heterogeneous shape, extent, and location are some of the factors that make these tumors difficult to resect, and in some cases inoperable. The amount of resected tumoris a factor also considered in longitudinal scans, when evaluating the apparent tumor for potential diagnosis of progression. Furthermore, there is mounting evidence that accurate segmentation of the various tumor sub-regions can offer the basis for quantitative image analysis towards prediction of patient overall survival. This study assesses thestate-of-the-art machine learning (ML) methods used for brain tumor image analysis in mpMRI scans, during the last seven instances of the International Brain Tumor Segmentation (BraTS) challenge, i.e., 2012-2018. Specifically, we focus on i) evaluating segmentations of the various glioma sub-regions in pre-operative mpMRI scans, ii) assessing potential tumor progression by virtue of longitudinal growth of tumor sub-regions, beyond use of the RECIST/RANO criteria, and iii) predicting the overall survival from pre-operative mpMRI scans of patients that underwent gross tota lresection. Finally, we investigate the challenge of identifying the best ML algorithms for each of these tasks, considering that apart from being diverse on each instance of the challenge, the multi-institutional mpMRI BraTS dataset has also been a continuously evolving/growing dataset.

Identifying the Best Machine Learning Algorithms for Brain Tumor Segmentation, Progression Assessment, and Overall Survival Prediction in the BRATS Challenge

Radiomics is an emerging field in quantitative imaging that uses advanced imaging features to objectively and quantitatively describe tumour phenotypes. Radiomic features have recently drawn considerable interest due to its potential predictive power for treatment outcomes and cancer genetics, which may have important applications in personalized medicine. In this technical review, we describe applications and challenges of the radiomic field. We will review radiomic application areas and technical issues, as well as proper practices for the designs of radiomic studies.

/pdf/applications-and-limitations-of-radiomics-38sh1zf0rc.pdf

Applications and limitations of radiomics

A novel region of interest (ROI) segmentation for detection of Glioblastoma multiforme (GBM) tumor in magnetic resonance (MR) images of the brain is proposed using a two-stage thresholding method. We have defined multiple intervals for multilevel thresholding using a novel meta-heuristic optimization technique called Discrete Curve Evolution. In each of these intervals, a threshold is selected by bi-level Otsu’s method. Then the ROI is extracted from only a single seed initialization, on the ROI, by the user. The proposed segmentation technique is more accurate as compared to the existing methods. Also the time complexity of our method is very low. The experimental evaluation is provided on contrast-enhanced T1-weighted MRI slices of three patients, having the corresponding ground truth of the tumor regions. The performance measure, based on Jaccard and Dice indices, of the segmented ROI demonstrated higher accuracy than existing methods.

ROI Segmentation from Brain MR Images with a Fast Multilevel Thresholding

Monitoring the land cover from space can help us in many ways such as understanding the natural phenomenon, human activities and managing the natural resources. The problem of classifying a satellite image-chip, into one class or multiple classes, is explored. These classes could be land cover, natural phenomenon or human activity. The classification problem is solved using multilabel classification framework. The used data is obtained from Kaggle, and explored an effective techniques inspired from transfer learning (which utilizes convolution neural network (CNN) as feature extractor) and a popular gradient boosting algorithm: XGBoost (as classifier). The performance of designed model is evaluated through the K-fold cross validation. The proposed technique gives F1-score of 0.87411, which is far better than state-of-the-art methods, like Random Forest and Gaussian Naive Bayes.

A novel multilabel classification of remote sensing images using XGBoost

A novel adaptive k-NN classifier for handling imbalance: Application to brain MRI

Change detection (CD) using remote sensing images have gained much attention in recent past due to its diverse applications. Devising reliable CD techniques that integrate huge topographical information is highly challenging. Researches in deep learning paradigm, particularly with Convolutional neural networks (CNN), have proven that CNN are efficient in abstracting knowledge from mul- tiple spectral bands, easy to be trained, and capable of deriving inference from unseen datasets. However, gathering training patterns are difficult in many real life problems and therefore, the pre-trained CNN models can be applied effectively. Hence, we consider three CNN models, VGG19, InceptionV3 and ResNet50 for feature extraction using transfer learning, followed by KMeans and Fuzzy C-Means(FCM) clustering algorithms for generating change maps. The proposed methods have been tested on two representative datasets of different land cover dynamics and have exhibited promising results with high overall accuracy and Kappa statistic (95.09 & 0.8173 respectively on Dubai city dataset and 97.12 & 0.8970 respectively on Texas dataset for Resnet+FCM) as well as superior to the state-of-the-art methods compared.

Unsupervised Change Detection in Remote Sensing Images Using CNN Based Transfer Learning

Glioblastoma Multiforme (GBM) is an aggressive form of glioma, exhibiting very poor survival. Genomic input, in the form of RNA sequencing data (RNA-seq), is expected to provide vital information about the characteristics of the genes that affect the Overall Survival (OS) of patients. This could have a significant impact on treatment planning. We present a new Autoencoder (AE)-based strategy for the prediction of survival (low or high) of GBM patients, using the RNA-seq data of 129 GBM samples from The Cancer Genome Atlas (TCGA). This is a novel interdisciplinary approach to integrating genomics with deep learning towards survival prediction. First, the Differentially Expressed Genes (DEGs) were selected using EdgeR. These were further reduced using correlation-based analysis. This was followed by the application of ranking with different feature subset selection and feature extraction algorithms, including the AE. In each case, fifty features were selected/extracted, for subsequent prediction with different classifiers. An exhaustive study for survival group prediction, using eight different classifiers with the accuracy and Area Under the Curve (AUC), established the superiority of the AE-based feature extraction method, called DeepSGP. It produced a very high accuracy (0.83) and AUC (0.90). Of the eight classifiers, using the extracted features by DeepSGP, the MLP was the best at Overall Survival (OS) prediction with an accuracy of 0.89 and an AUC of 0.97. The biological significance of the genes extracted by the AE were also analyzed to establish their importance. Finally, the statistical significance of the predicted output of the DeepSGP algorithm was established using the concordance index.

https://www.mdpi.com/2079-9292/10/12/1463/pdf

B. Uma Shankar

Papers

ROI Segmentation from Brain MR Images with a Fast Multilevel Thresholding

A novel multilabel classification of remote sensing images using XGBoost

A novel adaptive k-NN classifier for handling imbalance: Application to brain MRI

Unsupervised Change Detection in Remote Sensing Images Using CNN Based Transfer Learning

DeepSGP: Deep Learning for Gene Selection and Survival Group Prediction in Glioblastoma