Engineering Papers⌕ Search

Engineering topics

Karnowski, Thomas P.

Publications and source records attributed to Karnowski, Thomas P..

Database to Enable Facial Analysis for Driving Studies (DEFADS)

Naturalistic Driving Studies (NDS) collect and utilize data on drivers in real-world environments in instrumented vehicles. A common problem with such studies is driver privacy. In this work we collected a dataset of 77 human subjects performing scripted driving-related activities. We used three camera systems for the collection, including two high-resolution webcam devices as well as a third system from an actual NDS (the Second Strategic Highway Research Project, or SHRP2). This report covers the data collection process and summarizes the dataset, which will be made publicly available to researchers under a data usage license.

97 MATHEMATICS AND COMPUTING↗

Optimizing neural networks

A system and method design and optimize neural networks. The system and method include a data store that stores a plurality of gene vectors that represent diverse and distinct neural networks and an evaluation queue stored with the plurality of gene vectors. Secondary nodes construct, train, and evaluate the neural network and automatically render a plurality of fitness values asynchronously. A primary node executes a gene amplification on a select plurality of gene vectors, a crossing-over of the amplified gene vectors, and a mutation of the crossing-over gene vectors automatically and asynchronously, which are then transmitted to the evaluation queue. The process continuously repeats itself by processing the gene vectors inserted into the evaluation queue until a fitness level is reached, a network's accuracy level plateaus, a processing time period expires, or when some stopping condition or performance metric is met or exceeded.

97 MATHEMATICS AND COMPUTING↗

Evaluating Automated Face Identity-Masking Methods with Human Perception and a Deep Convolutional Neural Network

Face de-identification (or “masking”) algorithms have been developed in response to the prevalent use of video recordings in public places. Here, we evaluated the success of face identity masking for human perceivers and a deep convolutional neural network (DCNN). Eight de-identification algorithms were applied to videos of drivers’ faces, while they actively operated a motor vehicle. These masks were pre-selected to be applicable to low-quality video and to maintain coarse information about facial actions. Humans studied high-resolution images to learn driver identities and were tested on their recognition of active drivers in low-resolution videos. Faces in the videos were either unmasked or were masked by one of the eight algorithms. When participants were tested immediately after learning (Experiment 1), all masks reduced identification, with six of eight masks reducing identification to extremely poor performance. In a second experiment, two of the most effective masks were tested after a delay of 7 or 28 days. The delay did not further reduce identification of the masked faces. In all masked conditions, participants maintained stringent decision criteria, with low confidence in recognition, further indicating the effectiveness of the masks. Next, the DCNN performed an identity-matching task between high-resolution images and masked videos—a task analogous to that done by humans. The pattern of accuracy for the DCNN mirrored some, but not all, aspects of human performance, highlighting the need to test the effectiveness of identity masking for both humans and machines. The DCNN was also tested on its ability to match identity between masked and unmasked versions of the same video, based only on the face. DCNN performance for the eight masks offers insight into the nature of the information in faces that is coded in these networks.

97 MATHEMATICS AND COMPUTING↗

Reinforcement Learning-based Traffic Control to Optimize Energy Usage and Throughput (CRADA report)

The US roadways are critical to meeting the mobility and economic needs of the nation. The United States uses 28% of its energy in moving goods and people, with approximately 60% of that utilized by cars, light trucks, and motorcycles. Thus, improved transportation efficiency is vital to America’s economic progress. The increasing congestion and energy resource requirements of transportation systems for metropolitan areas require research in methods to improve and optimize control methods. Coordinating and optimizing traffic in urban areas may introduce hundreds of thousands of vehicles and traffic management systems, which can require high performance computing (HPC) resources to model and manage. In this work, we seek to use machine learning, computer vision, and HPC to improve the energy efficiency aspects of traffic control by leveraging GRIDSMART traffic cameras as sensors for adaptive traffic control, with a sensitivity to the fuel consumption characteristics of the traffic in the camera’s visual field. Traffic control use cases using reinforcement learning have been published and achieved good results. Surveys from DOE national laboratories estimate that the fuel cost of idling is 6 billion gallons wasted annually [Argonne National Laboratory, 2019]. GRIDSMART cameras—an existing, fielded commercial product—sense the presence of vehicles at intersections and replace more conventional sensors (such as inductive loops) to issue calls to traffic control. These cameras, which have horizon-to-horizon view, offer the potential for an improved view of the traffic environment which can be used to generate better control algorithms.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

The ORNL Overhead Vehicle Dataset (OOVD)

Roadways are critical to meeting the mobility and economic needs of the nation. The United States uses 28% of its energy in moving goods and people, with approximately 60% of that utilized by cars, light trucks, and motorcycles. Thus, improved transportation efficiency is vital to America’s economic progress. The increasing congestion and energy resource requirements of transportation systems for metropolitan areas require research in methods to improve and optimize control methods. Coordinating and optimizing traffic in urban areas may introduce hundreds of thousands of vehicles and traffic management systems, which can require high performance computing (HPC) resources to model and manage. This data set was created to understand the potential for machine learning, computer vision, and HPC to improve the energy efficiency aspects of traffic control by leveraging GRIDSMART traffic cameras as sensors for adaptive traffic control, with a sensitivity to the fuel consumption characteristics of the traffic in the camera’s visual field. GRIDSMART cameras—an existing, fielded commercial product—sense the presence of vehicles at intersections and replace more conventional sensors (such as inductive loops) to issue calls to traffic control. These cameras, which have horizon-to-horizon view, offer the potential for an improved view of the traffic environment which can be used to generate better control algorithms.

Karnowski, Thomas P.↗

167-PFlops Deep Learning for Electron Microscopy: From Learning Physics to Atomic Manipulation

An artificial intelligence system called MENNDL, which used 25,200 Nvidia Volta GPUs on Oak Ridge National Laboratory’s Summit machine, automatically designed an optimal deep learning network in order to extract structural information from raw atomic-resolution microscopy data. In a few hours, MENNDL creates and evaluates millions of networks using a scalable, parallel, asynchronous genetic algorithm augmented with a support vector machine to automatically find a superior deep learning network topology and hyper-parameter set than a human expert can find in months. For the application of electron microscopy, the system furthers the goal of improving our understanding of the electron-beam-matter interactions and real-time image-based feedback, which enables a huge step beyond human capacity toward nanofabricating materials automatically. Finally, MENNDL has been scaled to the 4,200 available nodes of Summit achieving a measured 152.5 PFlops, with an estimated sustained performance of 167 PFlops when the entire machine is available.

97 MATHEMATICS AND COMPUTING↗

A Dimensionally Aligned Signal Projection for Classification of Unintended Radiated Emissions

Characterization of unintended radiated emissions (URE) from electronic devices plays an important role in many research areas from electromagnetic interference to nonintrusive load monitoring to information system security. URE can provide insights for applications ranging from load disaggregation and energy efficiency to condition-based maintenance of equipment-based upon detected fault conditions. URE characterization often requires subject matter expertise to tailor transforms and feature extractors for the specific electrical devices of interest. We present a novel approach, named dimensionally aligned signal projection (DASP), for projecting aligned signal characteristics that are inherent to the physical implementation of many commercial electronic devices. These projections minimize the need for an intimate understanding of the underlying physical circuitry and significantly reduce the number of features required for signal classification. We present three possible DASP algorithms that leverage frequency harmonics, modulation alignments, and frequency peak spacings, along with a two-dimensional image manipulation method for statistical feature extraction. To demonstrate the ability of DASP to generate relevant features from URE, we measured the conducted URE from 14 residential electronic devices using a 2 MS/s collection system. Furthermore, a linear discriminant analysis classifier was trained using DASP generated features and was blind tested resulting in a greater than 90% classification accuracy for each of the DASP algorithms and an accuracy of 99.1% when DASP features are used in combination. Furthermore, we show that a rank reduced feature set of the combined DASP algorithms provides a 98.9% classification accuracy with only three features and outperforms a set of spectral features in terms of general classification as well as applicability across a broad number of devices.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Optimizing deep learning hyper-parameters through an evolutionary algorithm. In: MLHPC '15 Proceedings of the Workshop on Machine Learning in High-Performance Computing Environments, Article No. 4

There has been a recent surge of success in utilizing Deep Learning (DL) in imaging and speech applications for its relatively automatic feature generation and, in particular for convolutional neural networks (CNNs), high accuracy classification abilities. While these models learn their parameters through data-driven methods, model selection (as architecture construction) through hyper-parameter choices remains a tedious and highly intuition driven task. To address this, Multi-node Evolutionary Neural Networks for Deep Learning (MENNDL) is proposed as a method for automating network selection on computational clusters through hyper-parameter optimization performed via genetic algorithms.

Young, Steven R.↗

Method and system for the diagnosis of disease using retinal image content and an archive of diagnosed human patient data

A method for diagnosing diseases having retinal manifestations including retinal pathologies includes the steps of providing a CBIR system including an archive of stored digital retinal photography images and diagnosed patient data corresponding to the retinal photography images, the stored images each indexed in a CBIR database using a plurality of feature vectors, the feature vectors corresponding to distinct descriptive characteristics of the stored images. A query image of the retina of a patient is obtained. Using image processing, regions or structures in the query image are identified. The regions or structures are then described using the plurality of feature vectors. At least one relevant stored image from the archive based on similarity to the regions or structures is retrieved, and an eye disease or a disease having retinal manifestations in the patient is diagnosed based on the diagnosed patient data associated with the relevant stored image(s).

Tobin, Jr., Kenneth William↗

Retina Image Screening and Analysis Software Version 2.0

The software allows physicians or researchers to ground-truth images of retinas, identifying key physiological features and lesions that are indicative of disease. The software features methods to automatically detect the physiological features and lesions. The software contains code to measure the quality of images received from a telemedicine network; create and populate a database for a telemedicine network; review and report the diagnosis of a set of images; and also contains components to transmit images from a Zeiss camera to the network through SFTP.

Tobin, Jr., Kenneth W.↗