Digital Commons Network^™

Generation Of Phase Transitions Boundaries Via Convolutional Neural Networks, Christopher Alexis Ibarra

Open Access Theses & Dissertations

Accurate mapping of phase transitions boundaries is crucial in accurately modeling the equation of state of materials. The phase transitions can be structural (solid-solid) driven by temperature or pressure or a phase change like melting which defines the solid-liquid melt line. There exist many computational methods for evaluating the phase diagram at a particular point in temperature (T) and pressure (P). Most of these methods involve evaluation of a single (P,T) point at a time. The present work partially automates the search for phase boundaries lines utilizing a machine learning method based on convolutional neural networks and an efficient search …

How To Train Vision Transformer On Small-Scale Datasets?, Hanan Gani, Muzammal Naseer, Mohammad Yaqub

Computer Vision Faculty Publications

Vision Transformer (ViT), a radically different architecture than convolutional neural networks offers multiple advantages including design simplicity, robustness and state-of-the-art performance on many vision tasks. However, in contrast to convolutional neural networks, Vision Transformer lacks inherent inductive biases. Therefore, successful training of such models is mainly attributed to pre-training on large-scale datasets such as ImageNet with 1.2M or JFT with 300M images. This hinders the direct adaption of Vision Transformer for small-scale datasets. In this work, we show that self-supervised inductive biases can be learned directly from small-scale datasets and serve as an effective weight initialization scheme for fine-tuning. This …

Machine Learning For Aiding Blood Flow Velocity Estimation Based On Angiography, Swati Padhee, Mark Johnson, Hang Yi, Tanvi Banerjee, Zifeng Yang

Computer Science and Engineering Faculty Publications

Computational fluid dynamics (CFD) is widely employed to predict hemodynamic characteristics in arterial models, while not friendly to clinical applications due to the complexity of numerical simulations. Alternatively, this work proposed a framework to estimate hemodynamics in vessels based on angiography images using machine learning (ML) algorithms. First, the iodine contrast perfusion in blood was mimicked by a flow of dye diffusing into water in the experimentally validated CFD modeling. The generated projective images from simulations imitated the counterpart of light passing through the flow field as an analogy of X-ray imaging. Thus, the CFD simulation provides both the ground …

Medical Image Segmentation With Deep Convolutional Neural Networks, Chuanbo Wang

Theses and Dissertations

Medical imaging is the technique and process of creating visual representations of the body of a patient for clinical analysis and medical intervention. Healthcare professionals rely heavily on medical images and image documentation for proper diagnosis and treatment. However, manual interpretation and analysis of medical images are time-consuming, and inaccurate when the interpreter is not well-trained. Fully automatic segmentation of the region of interest from medical images has been researched for years to enhance the efficiency and accuracy of understanding such images. With the advance of deep learning, various neural network models have gained great success in semantic segmentation and …

Adversarial Pixel Restoration As A Pretext Task For Transferable Perturbations, Hashmat Shadab Malik, Shahina K. Kunhimon, Muzammal Nasser, Salman Khan, Fahad Shahbaz Khan

Computer Vision Faculty Publications

Transferable adversarial attacks optimize adversaries from a pretrained surrogate model and known label space to fool the unknown black-box models. Therefore, these attacks are restricted by the availability of an effective surrogate model. In this work, we relax this assumption and propose Adversarial Pixel Restoration as a self-supervised alternative to train an effective surrogate model from scratch under the condition of no labels and few data samples. Our training approach is based on a min-max objective which reduces overfitting via an adversarial objective and thus optimizes for a more generalizable surrogate model. Our proposed attack is complimentary to our adversarial …

Deep Convolution Neural Networks For Image Classification, Arun D. Kulkarni

Computer Science Faculty Publications and Presentations

Deep learning is a highly active area of research in machine learning community. Deep Convolutional Neural Networks (DCNNs) present a machine learning tool that enables the computer to learn from image samples and extract internal representations or properties underlying grouping or categories of the images. DCNNs have been used successfully for image classification, object recognition, image segmentation, and image retrieval tasks. DCNN models such as Alex Net, VGG Net, and Google Net have been used to classify large dataset having millions of images into thousand classes. In this paper, we present a brief review of DCNNs and results of our …

A New Automatic Bearing Fault Size Diagnosis Using Time-Frequency Images Of Cwt And Deep Transfer Learning Methods, Yilmaz Kaya, Fatma Kuncan, Hüseyi̇n Meti̇n Ertunç

Turkish Journal of Electrical Engineering and Computer Sciences

Bearings are generally used as bearings or turning elements. Bearings are subjected to high loads and rapid speeds. Furthermore, metal-to-metal contact within the bearing makes it sensitive. In today?s machines, bearing failures disrupt the operation of the system or completely stop the system. Bearing failures that can occur can cause enormous damage to the entire system. Therefore, it is necessary to anticipate bearing failures and to carry out a regular diagnostic examination. Various systems have been developed for fault diagnosis. In recent years, deep transfer learning (DTL) methods are often preferred in current bearing diagnosis models, as they provide time …

Edgenext: Efficiently Amalgamated Cnn-Transformer Architecture For Mobile Vision Applications, Muhammad Maaz, Abdelrahman Shaker, Hisham Cholakkal, Salman Khan, Syed Waqas Zamir, Rao Anwer, Fahad Shahbaz Khan

Computer Vision Faculty Publications

In the pursuit of achieving ever-increasing accuracy, large and complex neural networks are usually developed. Such models demand high computational resources and therefore cannot be deployed on edge devices. It is of great interest to build resource-efficient general purpose networks due to their usefulness in several application areas. In this work, we strive to effectively combine the strengths of both CNN and Transformer models and propose a new efficient hybrid architecture EdgeNeXt. Specifically in EdgeNeXt, we introduce split depth-wise transpose attention (SDTA) encoder that splits input tensors into multiple channel groups and utilizes depth-wise convolution along with self-attention across channel …

Learning Enriched Features For Fast Image Restoration And Enhancement, Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang, Ling Shao

Computer Vision Faculty Publications

Given a degraded input image, image restoration aims to recover the missing high-quality image content. Numerous applications demand effective image restoration, e.g., computational photography, surveillance, autonomous vehicles, and remote sensing. Significant advances in image restoration have been made in recent years, dominated by convolutional neural networks (CNNs). The widely-used CNN-based methods typically operate either on full-resolution or on progressively low-resolution representations. In the former case, spatial details are preserved but the contextual information cannot be precisely encoded. In the latter case, generated outputs are semantically reliable but spatially less accurate. This paper presents a new architecture with a holistic goal …

Malware Binary Image Classification Using Convolutional Neural Networks, John Kiger, Shen-Shyang Ho, Vahid Heydari

Faculty Scholarship for the College of Science & Mathematics

The persistent shortage of cybersecurity professionals combined with enterprise networks tasked with processing more data than ever before has led many cybersecurity experts to consider automating some of the most common and time-consuming security tasks using machine learning. One of these cybersecurity tasks where machine learning may prove advantageous is malware analysis and classification. To evade traditional detection techniques, malware developers are creating more complex malware. This is achieved through more advanced methods of code obfuscation and conducting more sophisticated attacks. This can make the manual process of analyzing malware an infinitely more complex task. Furthermore, the proliferation of malicious …

An Ensemble Approach For Patient Prognosis Of Head And Neck Tumor Using Multimodal Data, Numan Saeed, Roba Al Majzoub, Ikboljon Sobirov, Mohammad Yaqub

Computer Vision Faculty Publications

Accurate prognosis of a tumor can help doctors provide a proper course of treatment and, therefore, save the lives of many. Tradi-tional machine learning algorithms have been eminently useful in crafting prognostic models in the last few decades. Recently, deep learning algorithms have shown significant improvement when developing diag-nosis and prognosis solutions to different healthcare problems. However, most of these solutions rely solely on either imaging or clinical data. Utilizing patient tabular data such as demographics and patient med-ical history alongside imaging data in a multimodal approach to solve a prognosis task has started to gain more interest recently and …

A Survey Of Blind Modulation Classification Techniques For Ofdm Signals, Anand Kumar, Sudhan Majhi, Guan Gui, Hsiao-Chun Wu, Chau Yuen

Faculty Publications

Blind modulation classification (MC) is an integral part of designing an adaptive or intelligent transceiver for future wireless communications. Blind MC has several applications in the adaptive and automated systems of sixth generation (6G) communications to improve spectral efficiency and power efficiency, and reduce latency. It will become a integral part of intelligent software-defined radios (SDR) for future communication. In this paper, we provide various MC techniques for orthogonal frequency division multiplexing (OFDM) signals in a systematic way. We focus on the most widely used statistical and machine learning (ML) models and emphasize their advantages and limitations. The statistical-based blind …

Transformers In Vision: A Survey, Salman Khan, Muzammal Naseer, Munawar Hayat, Syed Waqas Zamir, Fahad Shahbaz Khan, Mubarak Shah

Computer Vision Faculty Publications

Astounding results from Transformer models on natural language tasks have intrigued the vision community to study their application to computer vision problems. Among their salient benefits, Transformers enable modeling long dependencies between input sequence elements and support parallel processing of sequence as compared to recurrent networks e.g., Long short-term memory (LSTM). Different from convolutional networks, Transformers require minimal inductive biases for their design and are naturally suited as set-functions. Furthermore, the straightforward design of Transformers allows processing multiple modalities (e.g., images, videos, text and speech) using similar processing blocks and demonstrates excellent scalability to very large capacity networks and huge …

Automatic Segmentation Of Head And Neck Tumor: How Powerful Transformers Are?, Ikboljon Sobirov, Otabek Nazarov, Hussain Alasmawi, Mohammad Yaqub

Computer Vision Faculty Publications

Cancer is one of the leading causes of death worldwide, and head and neck (H&N) cancer is amongst the most prevalent types. Positron emission tomography and computed tomography are used to detect and segment the tumor region. Clinically, tumor segmentation is extensively time-consuming and prone to error. Machine learning, and deep learning in particular, can assist to automate this process, yielding results as accurate as the results of a clinician. In this research study, we develop a vision transformers-based method to automatically delineate H&N tumor, and compare its results to leading convolutional neural network (CNN)-based models. We use multi-modal data …

Image-Data-Driven Deep Learning For Slope Stability Analysis, Behnam Azmoon

Dissertations, Master's Theses and Master's Reports

Landslides cause major infrastructural issues, damage the environment, and cause socio-economic disruptions. Therefore, various slope stability analysis methods have been developed to evaluate the stability of slopes and the probability of their failure. This dissertation attempts to take advantage of the recent advancements in remote sensing and computer technology to implement a deep-learning-based landslide prediction method.

Considering the novelty of this approach, this dissertation leads with proof-of-concept studies to evaluate and establish the suitability of deep learning models for slope stability analysis. To achieve this, a simulated 2D dataset of slope images was created with different geometries and soil properties. …

Cnn Based Sensor Fusion Method For Real-Time Autonomous Robotics Systems, Berat Yildiz, Aki̇f Durdu, Ahmet Kayabaşi, Mehmet Duramaz

Turkish Journal of Electrical Engineering and Computer Sciences

Autonomous robotic systems (ARS) serve in many areas of daily life. The sensors have critical importance for these systems. The sensor data obtained from the environment should be as accurate and reliable as possible and correctly interpreted by the autonomous robot. Since sensors have advantages and disadvantages over each other they should be used together to reduce errors. In this study, Convolutional Neural Network (CNN) based sensor fusion was applied to ARS to contribute the autonomous driving. In a real-time application, a camera and LIDAR sensor were tested with these networks. The novelty of this work is that the uniquely …

Construction Of A Repeatable Framework For Prostate Cancer Lesion Binary Semantic Segmentation Using Convolutional Neural Networks, Ian Vincent O. Mirasol, Patricia Angela R. Abu, Rosula Sj Reyes

Department of Information Systems & Computer Science Faculty Publications

Prostate cancer is the 3rd most diagnosed cancer overall. Current screening methods such as the prostate-specific antigen test could result in overdiagonosis and overtreatment while other methods such as a transrectal ultrasonography are invasive. Recent medical advancements have allowed the use of multiparametric MRI — a noninvasive and reliable screening process for prostate cancer. However, assessment would still vary from different professionals introducing subjectivity. While con-volutional neural network has been used in multiple studies to ob-jectively segment prostate lesions, due to the sensitivity of datasets and varying ground-truth established used in these studies, it is not possible to reproduce and …

Batch Normalization Preconditioning For Neural Network Training, Susanna Luisa Gertrude Lange

Theses and Dissertations--Mathematics

Batch normalization (BN) is a popular and ubiquitous method in deep learning that has been shown to decrease training time and improve generalization performance of neural networks. Despite its success, BN is not theoretically well understood. It is not suitable for use with very small mini-batch sizes or online learning. In this work, we propose a new method called Batch Normalization Preconditioning (BNP). Instead of applying normalization explicitly through a batch normalization layer as is done in BN, BNP applies normalization by conditioning the parameter gradients directly during training. This is designed to improve the Hessian matrix of the loss …

Facial Landmark Feature Fusion In Transfer Learning Of Child Facial Expressions, Megan A. Witherow, Manar D. Samad, Norou Diawara, Khan M. Iftekharuddin

Electrical & Computer Engineering Faculty Publications

Automatic classification of child facial expressions is challenging due to the scarcity of image samples with annotations. Transfer learning of deep convolutional neural networks (CNNs), pretrained on adult facial expressions, can be effectively finetuned for child facial expression classification using limited facial images of children. Recent work inspired by facial age estimation and age-invariant face recognition proposes a fusion of facial landmark features with deep representation learning to augment facial expression classification performance. We hypothesize that deep transfer learning of child facial expressions may also benefit from fusing facial landmark features. Our proposed model architecture integrates two input branches: a …

Deep Learning Based Superconducting Radio-Frequency Cavity Fault Classification At Jefferson Laboratory, Lasitha Vidyaratne, Adam Carpenter, Tom Powers, Chris Tennant, Khan M. Iftekharuddin, Md. Monibor Rahman, Anna S. Shabalina

Electrical & Computer Engineering Faculty Publications

This work investigates the efficacy of deep learning (DL) for classifying C100 superconducting radio-frequency (SRF) cavity faults in the Continuous Electron Beam Accelerator Facility (CEBAF) at Jefferson Lab. CEBAF is a large, high-power continuous wave recirculating linac that utilizes 418 SRF cavities to accelerate electrons up to 12 GeV. Recent upgrades to CEBAF include installation of 11 new cryomodules (88 cavities) equipped with a low-level RF system that records RF time-series data from each cavity at the onset of an RF failure. Typically, subject matter experts (SME) analyze this data to determine the fault type and identify the cavity of …