Data Driven Approach To Image Classification

With the availability of big image datasets and state-of-the-art computing hardware, data-driven machine learning approaches, particularly deep learning, have been used in numerous medical image (CT-scans, MRI, PET, SPECT, etc..) computing tasks, ranging from image reconstruction, super-resolution, segmentation, registration all the way to disease classification and survival prediction. However, training such high-precision approaches often require large amounts of data to be collected and labelled and high-capacity graphics processing units (GPUs) installed, which are resource intensive and hence not always practical. Other hurdles such as the generalization ability to unseen new data and difficulty to interpret and explain can prevent their deployment to those clinical applications which deem such abilities imperative.

Content-Based Image Classification: Efficient Machine Learning Using Robust Feature Extraction Techniques is a comprehensive guide to research with invaluable image data. Social Science Research Network has revealed that 65% of people are visual learners. Research data provided by Hyerle (2000) has clearly shown 90% of information in the human brain is visual. Thus, it is no wonder that visual information processing in the brain is 60,000 times faster than text-based information (3M Corporation, 2001). Recently, we have witnessed a significant surge in conversing with images due to the popularity of social networking platforms. The other reason for embracing usage of image data is the mass availability of high-resolution cellphone cameras. Wide usage of image data in diversified application areas including medical science, media, sports, remote sensing, and so on, has spurred the need for further research in optimizing archival, maintenance, and retrieval of appropriate image content to leverage data-driven decision-making. This book demonstrates several techniques of image processing to represent image data in a desired format for information identification. It discusses the application of machine learning and deep learning for identifying and categorizing appropriate image data helpful in designing automated decision support systems. The book offers comprehensive coverage of the most essential topics, including: Image feature extraction with novel handcrafted techniques (traditional feature extraction) Image feature extraction with automated techniques (representation learning with CNNs) Significance of fusion-based approaches in enhancing classification accuracy MATLAB® codes for implementing the techniques Use of the Open Access data mining tool WEKA for multiple tasks The book is intended for budding researchers, technocrats, engineering students, and machine learning/deep learning enthusiasts who are willing to start their computer vision journey with content-based image recognition. The readers will get a clear picture of the essentials for transforming the image data into valuable means for insight generation. Readers will learn coding techniques necessary to propose novel mechanisms and disruptive approaches. The WEKA guide provided is beneficial for those uncomfortable coding for machine learning algorithms. The WEKA tool assists the learner in implementing machine learning algorithms with the click of a button. Thus, this book will be a stepping-stone for your machine learning journey. Please visit the author's website for any further guidance at https://www.rikdas.com/

Image classification has been a core topic in the computer vision community. Its recent success with convolutional neural network (CNN) algorithm has led to various real world applications such as large scale management of photos/videos on cloud/social-media, image based search for online retailers, self-driving cars, building robots and healthcare. Image classification can be broadly categorized into binary, multi-class and multi-label classification problems. Binary classification involves assigning one of the two class labels to an instance. In multi-class classification problem, an instance should be categorized into one of more than two classes. Multi-label classification is a generalized version of the multi-class classification problem where each image is assigned multiple labels as opposed to a single label. In this work, we first present various methods that take advantage of deep representations (fully connected layer of pre-trained CNN on the ImageNet dataset) and yield better performance on multi-label classification when compared to methods that use over a dozen conventional visual features. Following the success of deep representations, we intend to build a generic end-to-end deep learning framework to address all three problem categories of image classification. However, there are still no well established guidelines (in terms of choosing the number of layers to go deeper, the number of kernels and the size, the type of regularizer, the choice of non-linear function, etc.) to build an efficient deep neural network and often network architecture design is specific to a problem/dataset. Hence, we present some initial efforts in building a computational framework called Deep Decision Network (DDN) which is completely data-driven. DDN is a tree-like structured built stage-wise. During the learning phase, starting from the root network node, DDN automatically builds a network that splits the data into disjoint clusters of classes which would be handled by the subsequent expert networks. This results in a tree-like structured network driven by the data. The proposed approach provides an insight into the data by identifying the group of classes that are hard to classify and require more attention when compared to others. This feature is crucial for people trying to solve the problem with little or no domain knowledge, especially for applications in medical domain. Initially, we evaluate DDN on a binary classification problem and later extend it to more challenging multi-class and multi-label classification problems. The extension of DDN to multi-class and multi-label involves some changes but they still operate under the same underlying principle. In all the three cases, the proposed approach is tested for its recognition performance and scalability on publicly available datasets providing comparison to other methods.

The book presents current research advances, both academic and industrial, in machine learning, artificial intelligence, and data analytics for biomedical and healthcare applications. The book deals with key challenges associated with biomedical data analysis including higher dimensions, class imbalances, smaller database sizes, etc. It also highlights development of novel pattern recognition and machine learning methods specific to medical and genomic data, which is extremely necessary but highly challenging. The book will be useful for healthcare professionals who have access to interesting data sources but lack the expertise to use data mining effectively.

Typically, inverse imaging problems deal with the reconstruction of images from the sensor measurements where sensors can take form of any imaging modality like camera, radar, hyperspectral or medical imaging systems. In an ideal scenario, we can reconstruct the images via applying an inversion procedure from these sensors’ measurements, but practical applications have several challenges: the measurement acquisition process is heavily corrupted by the noise, the forward model is not exactly known, and non-linearities or unknown physics of the data acquisition play roles. Hence, perfect inverse function is not exactly known for immaculate image reconstruction. To this end, in this dissertation, I propose an automatic sensing and reconstruction scheme based on deep learning within the compressive sensing (CS) framework to solve the computational imaging problems. Here, I develop a data-driven approach to learn both the measurement matrix and the inverse reconstruction scheme for a given class of signals, such as images. This approach paves the way for end-to-end learning and reconstruction of signals with the aid of cascaded fully connected and multistage convolutional layers with a weighted loss function in an adversarial learning framework. I also propose to extend our analysis to introduce data driven models to directly classify from compressed measurements through joint reconstruction and classification. I develop constrained measurement learning framework and demonstrate higher performance of the proposed approach in the field of typical image reconstruction and hyperspectral image classification tasks. Finally, I also propose a single data driven network that can take and reconstruct images at multiple rates of signal acquisition. In summary, this dissertation proposes novel methods on the data driven measurement acquisition for sparse signal reconstruction and classification, learning measurements for given constraints underlying the requirement of the hardware for different applications, and producing a common data driven platform for learning measurements to reconstruct signals at multiple rates. This dissertation opens the path to the learned sensing systems. The future research can use these proposed data driven approaches as the pivotal factors to accomplish task-specific smart sensors in several real-world applications.

This book constitutes the refereed proceedings of the Third International Conference on Dynamic Data Driven Application Systems, DDDAS 2020, held in Boston, MA, USA, in October 2020. The 21 full papers and 14 short papers presented in this volume were carefully reviewed and selected from 40 submissions. They cover topics such as: digital twins; environment cognizant adaptive-planning systems; energy systems; materials systems; physics-based systems analysis; imaging methods and systems; and learning systems.

This book is a compilation of peer-reviewed papers presented at the International Conference on Machine Intelligence and Data Science Applications, organized by the School of Computer Science, University of Petroleum & Energy Studies, Dehradun, India, during 4–5 September 2020. The book addresses the algorithmic aspect of machine intelligence which includes the framework and optimization of various states of algorithms. Variety of papers related to wide applications in various fields like data-driven industrial IoT, bioinformatics, network and security, autonomous computing and various other aligned areas. The book concludes with interdisciplinary applications like legal, health care, smart society, cyber-physical system and smart agriculture. All papers have been carefully reviewed. The book is of interest to computer science engineers, lecturers/researchers in machine intelligence discipline and engineering graduates.

A textbook covering data-science and machine learning methods for modelling and control in engineering and science, with Python and MATLAB®.

Phishing Detection Using Content-Based Image Classification is an invaluable resource for any deep learning and cybersecurity professional and scholar trying to solve various cybersecurity tasks using new age technologies like Deep Learning and Computer Vision. With various rule-based phishing detection techniques at play which can be bypassed by phishers, this book provides a step-by-step approach to solve this problem using Computer Vision and Deep Learning techniques with significant accuracy. The book offers comprehensive coverage of the most essential topics, including: Programmatically reading and manipulating image data Extracting relevant features from images Building statistical models using image features Using state-of-the-art Deep Learning models for feature extraction Build a robust phishing detection tool even with less data Dimensionality reduction techniques Class imbalance treatment Feature Fusion techniques Building performance metrics for multi-class classification task Another unique aspect of this book is it comes with a completely reproducible code base developed by the author and shared via python notebooks for quick launch and running capabilities. They can be leveraged for further enhancing the provided models using new advancement in the field of computer vision and more advanced algorithms.