Engineering Commons^™

An Analysis Of Precision: Occlusion And Perspective Geometry’S Role In 6d Pose Estimation, Jeffrey Choate, Derek Worth, Scott Nykl, Clark N. Taylor, Brett J. Borghetti, Christine M. Schubert Kabban

Faculty Publications

Achieving precise 6 degrees of freedom (6D) pose estimation of rigid objects from color images is a critical challenge with wide-ranging applications in robotics and close-contact aircraft operations. This study investigates key techniques in the application of YOLOv5 object detection convolutional neural network (CNN) for 6D pose localization of aircraft using only color imagery. Traditional object detection labeling methods suffer from inaccuracies due to perspective geometry and being limited to visible key points. This research demonstrates that with precise labeling, a CNN can predict object features with near-pixel accuracy, effectively learning the distinct appearance of the object due to perspective …

Enhanced Two-Step Deep-Learning Approach For Electromagnetic-Inverse-Scattering Problems: Frequency Extrapolation And Scatterer Reconstruction, Huan Huan Zhang, He Ming Yao, Lijun Jiang, Michael Ng

Electrical and Computer Engineering Faculty Research & Creative Works

The electromagnetic-inverse-scattering (EMIS) problem is solved by a novel two-step deep-learning (DL) approach in this article. The newly proposed two-step DL approach not only predicts the multifrequency EM scattered field, but also overcomes the limitation of the conventional methods for solving EMIS problems, such as expensive computational cost, strong ill-conditions, and invalidity on high contrast. In the first step, the complex-valued deep residual convolutional neural network (DRCNN) is utilized to predict multifrequency EM scattered fields only using single-frequency EM scattered field information. Based on a new complex-valued deep convolutional encoder-decoder (DCED) structure, the second step utilizes the obtained multifrequency EM …

Implementing The Fast Full-Wave Electromagnetic Forward Solver Using The Deep Convolutional Encoder-Decoder Architecture, He Ming Yao, Lijun Jiang, Michael Ng

Electrical and Computer Engineering Faculty Research & Creative Works

In this communication, a novel deep learning (DL)-based solver is proposed for the electromagnetic forward (EMF) process. It is based on the complex-valued deep convolutional neural networks (DConvNets) comprising an encoder network and a corresponding decoder network with pixel-wise regression layer. The encoder network takes the incident EM wave and the contrast (permittivity) distribution of the object as the input. It channels the processed data into the corresponding decoder network to predict the total EM field due to the scatter of the input incident EM wave. The training of the proposed DConvNets is done using the simple synthetic dataset. Due …

Enhanced Supervised Descent Learning Technique For Electromagnetic Inverse Scattering Problems By The Deep Convolutional Neural Networks, He Ming Yao, Rui Guo, Maokun Li, Lijun Jiang, Michael Kwok Po Ng

Electrical and Computer Engineering Faculty Research & Creative Works

This work proposes a novel deep learning (DL) framework to solve the electromagnetic inverse scattering (EMIS) problems. The proposed framework integrates the complex-valued deep convolutional neural network (DConvNet) into the supervised descent method (SDM) to realize both off-line training and on-line 'imaging' prediction for EMIS. The offline training consists of two parts: 1) DConvNet training: the training dataset is created, and the proposed DConvNet is trained to realize the EM forward process and 2) SDM training: the trained DConvNet is integrated into the SDM framework, and the average descent directions between the initial prediction and the true label of SDM …

Ad-Corre: Adaptive Correlation-Based Loss For Facial Expression Recognition In The Wild, Ali Pourramezan Fard, Mohammad H. Mahoor

Electrical and Computer Engineering: Faculty Scholarship

Automated Facial Expression Recognition (FER) in the wild using deep neural networks is still challenging due to intra-class variations and inter-class similarities in facial images. Deep Metric Learning (DML) is among the widely used methods to deal with these issues by improving the discriminative power of the learned embedded features. This paper proposes an Adaptive Correlation (Ad-Corre) Loss to guide the network towards generating embedded feature vectors with high correlation for within-class samples and less correlation for between-class samples. Ad-Corre consists of 3 components called Feature Discriminator, Mean Discriminator, and Embedding Discriminator. We design the Feature Discriminator component to guide …

Heuristic-Based Automatic Pruning Of Deep Neural Networks, Tejalal Choudhary, Vipul Mishra, Anurag Goswami, Jagannathan Sarangapani

Electrical and Computer Engineering Faculty Research & Creative Works

The performance of a deep neural network (deep NN) is dependent upon a significant number of weight parameters that need to be trained which is a computational bottleneck. The growing trend of deeper architectures poses a restriction on the training and inference scheme on resource-constrained devices. Pruning is an important method for removing the deep NN's unimportant parameters and making their deployment easier on resource-constrained devices for practical applications. In this paper, we proposed a heuristics-based novel filter pruning method to automatically identify and prune the unimportant filters and make the inference process faster on devices with limited resource availability. …

Resampling And Super-Resolution Of Hexagonally Sampled Images Using Deep Learning, Dylan Flaute, Russell C. Hardie, Hamed Elwarfalli

Electrical and Computer Engineering Faculty Publications

Super-resolution (SR) aims to increase the resolution of imagery. Applications include security, medical imaging, and object recognition. We propose a deep learning-based SR system that takes a hexagonally sampled low-resolution image as an input and generates a rectangularly sampled SR image as an output. For training and testing, we use a realistic observation model that includes optical degradation from diffraction and sensor degradation from detector integration. Our SR approach first uses non-uniform interpolation to partially upsample the observed hexagonal imagery and convert it to a rectangular grid. We then leverage a state-of-the-art convolutional neural network (CNN) architecture designed for SR …

Deep Learning For Anisoplanatic Optical Turbulence Mitigation In Long-Range Imaging, Matthew A. Hoffmire, Russell C. Hardie, Michael A. Rucci, Richard Van Hook, Barry K. Karch

Electrical and Computer Engineering Faculty Publications

We present a deep learning approach for restoring images degraded by atmospheric optical turbulence. We consider the case of terrestrial imaging over long ranges with a wide field-of-view. This produces an anisoplanatic imaging scenario where turbulence warping and blurring vary spatially across the image. The proposed turbulence mitigation (TM) method assumes that a sequence of short-exposure images is acquired. A block matching (BM) registration algorithm is applied to the observed frames for dewarping, and the resulting images are averaged. A convolutional neural network (CNN) is then employed to perform spatially adaptive restoration. We refer to the proposed TM algorithm as …

Fifnet: A Convolutional Neural Network For Motion-Based Multiframe Super-Resolution Using Fusion Of Interpolated Frames, Hamed Elwarfalli, Russell C. Hardie

Electrical and Computer Engineering Faculty Publications

We present a novel motion-based multiframe image super-resolution (SR) algorithm using a convolutional neural network (CNN) that fuses multiple interpolated input frames to produce an SR output. We refer to the proposed CNN and associated preprocessing as the Fusion of Interpolated Frames Network (FIFNET). We believe this is the first such CNN approach in the literature to perform motion-based multiframe SR by fusing multiple input frames in a single network. We study the FIFNET using translational interframe motion with both fixed and random frame shifts. The input to the network is a sequence of interpolated and aligned frames. One key …

Predicting Residential Energy Consumption Using Wavelet Decomposition With Deep Neural Network, Dagimawi Eneyew, Miriam A M Capretz, Girma Bitsuamlak, London Hydro

Electrical and Computer Engineering Publications

Electricity consumption is accelerating due to economic and population growth. Hence, energy consumption prediction is becoming vital for overall consumption management and infrastructure planning. Recent advances in smart electric meter technology are making high-resolution energy consumption data available. However, many parameters influencing energy consumption are not typically monitored for residential buildings. Therefore, this study’s main objective is to develop a data-driven energy consumption forecasting model (next-hour consumption) for residential houses solely based on analyzing electricity consumption data. This research proposes a deep neural network architecture that combines stationary wavelet transform features and convolutional neural networks. The proposed approach utilizes automatically …

Enhanced Deep Learning Approach Based On The Deep Convolutional Encoder-Decoder Architecture For Electromagnetic Inverse Scattering Problems, He Ming Yao, Lijun Jiang, Wei E.I. Sha

Electrical and Computer Engineering Faculty Research & Creative Works

This letter proposes a novel deep learning (DL) approach to resolve the electromagnetic inverse scattering (EMIS) problems. The conventional approaches of resolving EMIS problems encounter assorted difficulties, such as high contrast, high computational cost, inevitable intrinsic nonlinearity, and strong ill-posedness. To surmount these difficulties, a novel DL approach is proposed based on a novel complex-valued deep fully convolutional neural network structure. The proposed complex-valued deep learning model for solving EMIS problems composes of an encoder network and its corresponding decoder network, followed by a final pixel-wise regression layer. The complex-valued encoder network extracts feature fragments from received scattered field data, …

Deep Learning With Context Encoding For Semantic Brain Tumor Segmentation And Patient Survival Prediction, Linmin Pei, Lasitha Vidyaratne, Md Monibor Rahman, Khan M. Iftekharuddin

Electrical & Computer Engineering Faculty Publications

One of the most challenging problems encountered in deep learning-based brain tumor segmentation models is the misclassification of tumor tissue classes due to the inherent imbalance in the class representation. Consequently, strong regularization methods are typically considered when training large-scale deep learning models for brain tumor segmentation to overcome undue bias towards representative tissue types. However, these regularization methods tend to be computationally exhaustive, and may not guarantee the learning of features representing all tumor tissue types that exist in the input MRI examples. Recent work in context encoding with deep CNN models have shown promise for semantic segmentation of …

Two-Step Enhanced Deep Learning Approach For Electromagnetic Inverse Scattering Problems, He Ming Yao, Wei E.I. Sha, Lijun Jiang

Electrical and Computer Engineering Faculty Research & Creative Works

In this letter, a new deep learning (DL) approach is proposed to solve the electromagnetic inverse scattering (EMIS) problems. The conventional methods for solving inverse problems face various challenges including strong ill-conditions, high contrast, expensive computation cost, and unavoidable intrinsic nonlinearity. To overcome these issues, we propose a new two-step machine learning based approach. In the first step, a complex-valued deep convolutional neural network is employed to retrieve initial contrasts (permittivity's) of dielectric scatterers from measured scattering data. In the second step, the previously obtained contrasts are input into a complex-valued deep residual convolutional neural network to refine the reconstruction …

Source Reconstruction Method Based On Machine Learning Algorithms, He Ming Yao, Lijun Jiang, Wei E.I. Sha

Electrical and Computer Engineering Faculty Research & Creative Works

This paper proposes a new source reconstruction method (SRM) based on deep learning. The conventional SRM usually requires oversampled measurements data to ensure higher accuracy. Thus, conventional SRM numerical system is usually highly singular. A deep convolutional neural network (ConvNet) is proposed to reconstruct the equivalent sources of the target to overcome difficulty. The deep ConvNet allows us to employ less data samples. Besides, the ill-conditioned numerical system can be effectively avoided. Numerical examples are presented to demonstrate the feasibility and accuracy of the proposed method. Its performance is also compared with the traditional neural network and interpolation method. Moreover, …

Applying Deep Learning Approach To The Far-Field Subwavelength Imaging Based On Near-Field Resonant Metalens At Microwave Frequencies, He Ming Yao, Min Li, Lijun Jiang

Electrical and Computer Engineering Faculty Research & Creative Works

In this paper, we utilize the deep learning approach for the subwavelength imaging in far-field, which is realized by the near-field resonant metalens at microwave frequencies. The resonating metalens consisting of split-ring resonators (SRRs) are equipped with the strong magnetic coupling ability and can convert evanescent waves into propagating waves using the localized resonant modes. The propagating waves in the far-field are utilized as the input of a trained deep convolutional neural network (CNN) to realize the imaging. The training data for establishing the deep CNN are obtained by the EM simulation tool. Besides, the white Gaussian noise is added …

Deep Learning Nuclei Detection In Digitized Histology Images By Superpixels, Sudhir Sornapudi, R. Joe Stanley, William V. Stoecker, Haidar Almubarak, Rodney Long, Sameer Antani, George Thoma, Rosemary Zuna, Shelliane R. Frazier

Electrical and Computer Engineering Faculty Research & Creative Works

Background: Advances in image analysis and computational techniques have facilitated automatic detection of critical features in histopathology images. Detection of nuclei is critical for squamous epithelium cervical intraepithelial neoplasia (CIN) classification into normal, CIN1, CIN2, and CIN3 grades.

Methods: In this study, a deep learning (DL)-based nuclei segmentation approach is investigated based on gathering localized information through the generation of superpixels using a simple linear iterative clustering algorithm and training with a convolutional neural network.

Results: The proposed approach was evaluated on a dataset of 133 digitized histology images and achieved an overall nuclei detection (object-based) accuracy of 95.97%, with …

Convolutional Neural Network Based Localized Classification Of Uterine Cervical Cancer Digital Histology Images, Haidar A. Almubarak, R. Joe Stanley, Rodney Long, Sameer Antani, George Thoma, Rosemary Zuna, Shelliane R. Frazier

Electrical and Computer Engineering Faculty Research & Creative Works

In previous research, we introduced an automated localized, fusion-based algorithm to classify squamous epithelium into Normal, CIN1, CIN2, and CIN3 grades of cervical intraepithelial neoplasia (CIN). The approach partitioned the epithelium into 10 segments. Image processing and machine vision algorithms were used to extract features from each segment. The features were then used to classify the segment and the result was fused to classify the whole epithelium. This research extends the previous research by dividing each of the 10 segments into 3 parts and uses a convolutional neural network to classify the 3 parts. The result is then fused to …