Search NASA⌕ Search

SEARCH · Search NASA

Results for “feature space”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records

Tensorized Feature Spaces for Feature Explosion

In this paper 1 1 This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-000R22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan) This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-000R22725., we present a novel framework that uses tensor factorization to generate richer feature spaces for pixel classification in hyperspectral images. In particular, we assess the performance of different tensor rank decomposition methods as compared to the traditional kernel-based approaches for the hyperspectral image classification problem. We propose Orion, which takes as input a hyperspectral image tensor and a rank and outputs an enhanced feature space from the factor matrices of the decomposed tensor. Our method is a feature explosion technique that inherently maps low dimensional input space in $\mathbb{R}^{K}$ to high dimensional space in $\mathbb{R}^{R}$ , where $R\gg K$ , say in the order of 1000x, like a kernel. We show how the proposed method exploits the multi-linear structure of hyperspectral three dimensional tensor. We demonstrate the effectiveness of our method with experiments on three publicly available hyperspectral datasets with labeled pixels and compare their classification performance against traditional linear and non-linear supervised learning methods such as SVM with Linear, Polynomial, RBF kernels, and the Multi-Layer Perceptron model. Finally, we explore the relationship between the rank of the tensor decomposition and the classification accuracy using several hyperspectral datasets with ground truth.

Pasricha, Ravdeep Singh↗

Trajectory analysis via a geometric feature space approach

This study aimed to organize a body of trajectories in order to identify, search for and classify both common and uncommon behaviors among objects such as aircraft and ships. Existing comparison functions such as the Fréchet distance are computationally expensive and yield counterintuitive results in some cases. We propose an approach using feature vectors whose components represent succinctly the salient information in trajectories. These features incorporate basic information such as the total distance traveled and the distance between start/stop points as well as geometric features related to the properties of the convex hull, trajectory curvature and general distance geometry. Additionally, these features can generally be mapped easily to behaviors of interest to humans who are searching large databases. Most of these geometric features are invariant under rigid transformation. Furthermore, we demonstrate the use of different subsets of these features to identify trajectories similar to an exemplar, cluster a database of several hundred thousand trajectories and identify outliers.

97 MATHEMATICS AND COMPUTING↗

TopFusion: Using Topological Feature Space for Fusion and Imputation in Multi-Modal Data

We present a novel multi-modal data fusion technique using topological features. The method, TopFusion, leverages the flexibility of topological data analysis tools (namely persistent homology and persistence images) to map multi-modal datasets into a common feature space by forming a new multi-channel persistence image. Each channel in the image is representative of a view of the data from a modality-dependent filtration. We demonstrate that the topological perspective we take allows for more effective data reconstruction, i.e. imputation. In particular, by performing imputation in topological feature space we are able to outperform the same imputation techniques applied to raw data or alternatively derived features. We show that TopFusion representations can be used as input to downstream deep learning-based computer vision models and doing so achieves comparable performance to other fusion methods for classification on two multi-modal datasets.

Myers, Audun D.↗

Complete fold annotation of the human proteome using a novel structural feature space

Recognition of protein structural fold is the starting point for many structure prediction tools and protein function inference. Fold prediction is computationally demanding and recognizing novel folds is difficult such that the majority of proteins have not been annotated for fold classification. Here we describe a new machine learning approach using a novel feature space that can be used for accurate recognition of all 1,221 currently known folds and inference of unknown novel folds. We show that our method achieves better than 94% accuracy even when many folds have only one training example. We demonstrate the utility of this method by predicting the folds of 34,330 human protein domains and showing that these predictions can yield useful insights into potential biological function, such as prediction of RNA-binding ability. Finally, our method can be applied to de novo fold prediction of entire proteomes and identify candidate novel fold families.

59 BASIC BIOLOGICAL SCIENCES↗

Block segmentation in feature space for realtime object detection in high granularity images

Computer vision has applications in object detection, image recognition and classification, and object tracking. One of the challenges of computer vision is the presence of useful information at multiple distance scales. Filtering techniques may sacrifice details at small scales in order to prioritize the analysis of large-scale features of the image. We present a strategy for coarse-graining multidimensional data while maintaining fine-grained detail for subsequent analysis. The algorithm is based on fixed-size block segmentation in the feature space. We apply this strategy to solve the long-standing challenge of detecting particle trajectories at the Large Hadron Collider in real time.

Computer vision↗

Feature-space assessment of electrical impedance tomography coregistered with computed tomography in detecting multiple contrast targets

Purpose: Fusion of electrical impedance tomography (EIT) with computed tomography (CT) can be useful as a clinical tool for providing additional physiological information about tissues, but requires suitable fusion algorithms and validation procedures. This work explores the feasibility of fusing EIT and CT images using an algorithm for coregistration. The imaging performance is validated through feature space assessment on phantom contrast targets. Methods: EIT data were acquired by scanning a phantom using a circuit, configured for injecting current through 16 electrodes, placed around the phantom. A conductivity image of the phantom was obtained from the data using electrical impedance and diffuse optical tomography reconstruction software (EIDORS). A CT image of the phantom was also acquired. The EIT and CT images were fused using a region of interest (ROI) coregistration fusion algorithm. Phantom imaging experiments were carried out on objects of different contrasts, sizes, and positions. The conductive medium of the phantoms was made of a tissue-mimicking bolus material that is routinely used in clinical radiation therapy settings. To validate the imaging performance in detecting different contrasts, the ROI of the phantom was filled with distilled water and normal saline. Spatially separated cylindrical objects of different sizes were used for validating the imaging performance in multiple target detection. Analyses of the CT, EIT and the EIT/CT phantom images were carried out based on the variations of contrast, correlation, energy, and homogeneity, using a gray level co-occurrence matrix (GLCM). A reference image of the phantom was simulated using EIDORS, and the performances of the CT and EIT imaging systems were evaluated and compared against the performance of the EIT/CT system using various feature metrics, detectability, and structural similarity index measures. Results: In detecting distilled and normal saline water in bolus medium, EIT as a stand-alone imaging system showed contrast discrimination of 47%, while the CT imaging system showed a discrimination of only 1.5%. The structural similarity index measure showed a drop of 24% with EIT imaging compared to CT imaging. The average detectability measure for CT imaging was found to be 2.375 ± 0.19 before fusion. After complementing with EIT information, the detectability measure increased to 11.06 ± 2.04. Based on the feature metrics, the functional imaging quality of CT and EIT were found to be 2.29% and 86%, respectively, before fusion. Structural imaging quality was found to be 66% for CT and 16% for EIT. After fusion, functional imaging quality improved in CT imaging from 2.29% to 42% and the structural imaging quality of EIT imaging changed from 16% to 66%. The improvement in image quality was also observed in detecting objects of different sizes. Conclusions: The authors found a significant improvement in the contrast detectability performance of CT imaging when complemented with functional imaging information from EIT. Along with the feature assessment metrics, the concept of complementing CT with EIT imaging can lead to an EIT/CT imaging modality which might fully utilize the functional imaging abilities of EIT imaging, thereby enhancing the quality of care in the areas of cancer diagnosis and radiotherapy treatment planning.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Characterization of masses in digital breast tomosynthesis: Comparison of machine learning in projection views and reconstructed slices

Purpose: In digital breast tomosynthesis (DBT), quasi-three-dimensional (3D) structural information is reconstructed from a small number of 2D projection view (PV) mammograms acquired over a limited angular range. The authors developed preliminary computer-aided diagnosis (CADx) methods for classification of malignant and benign masses and compared the effectiveness of analyzing lesion characteristics in the reconstructed DBT slices and in the PVs. Methods: A data set of MLO view DBT of 99 patients containing 107 masses (56 malignant and 51 benign) was collected at the Massachusetts General Hospital with IRB approval. The DBTs were obtained with a GE prototype system which acquired 11 PVs over a 50 deg. arc. The authors reconstructed the DBTs at 1 mm slice interval using a simultaneous algebraic reconstruction technique. The region of interest (ROI) containing the mass was marked by a radiologist in the DBT volume and the corresponding ROIs on the PVs were derived based on the imaging geometry. The subsequent processes were fully automated. For classification of masses using the DBT-slice approach, the mass on each slice was segmented by an active contour model initialized with adaptive k-means clustering. A spiculation likelihood map was generated by analysis of the gradient directions around the mass margin and spiculation features were extracted from the map. The rubber band straightening transform (RBST) was applied to a band of pixels around the segmented mass boundary. The RBST image was enhanced by Sobel filtering in the horizontal and vertical directions, from which run-length statistics texture features were extracted. Morphological features including those from the normalized radial length were designed to describe the mass shape. A feature space composed of the spiculation features, texture features, and morphological features extracted from the central slice alone and seven feature spaces obtained by averaging the corresponding features from three to 19 slices centered at the central slice were compared. For classification of masses using the PV approach, a feature extraction process similar to that described above for the DBT approach was performed on the ROIs from the individual PVs. Six feature spaces obtained from the central PV alone and by averaging the corresponding features from three to 11 PVs were formed. In each feature space for either the DBT-slice or the PV approach, a linear discriminant analysis classifier with stepwise feature selection was trained and tested using a two-loop leave-one-case-out resampling procedure. Simplex optimization was used to guide feature selection automatically within the training set in each leave-one-case-out cycle. The performance of the classifiers was evaluated by the area (A{sub z}) under the receiver operating characteristic curve. Results: The test A{sub z} values from the DBT-slice approach ranged from 0.87{+-}0.03 to 0.93{+-}0.02, while those from the PV approach ranged from 0.78{+-}0.04 to 0.84{+-}0.04. The highest test A{sub z} of 0.93{+-}0.02 from the nine-DBT-slice feature space was significantly (p=0.006) better than the highest test A{sub z} of 0.84{+-}0.04 from the nine-PV feature space. Conclusion: The features of breast lesions extracted from the DBT slices consistently provided higher classification accuracy than those extracted from the PV images.

60 APPLIED LIFE SCIENCES↗

System-independent characterization of materials using dual-energy computed tomography

In this study, we present a new decomposition approach for dual-energy computed tomography (DECT) called SIRZ that provides precise and accurate material description, independent of the scanner, over diagnostic energy ranges (30 to 200 keV). System independence is achieved by explicitly including a scanner-specific spectral description in the decomposition method, and a new X-ray-relevant feature space. The feature space consists of electron density, ρ e , and a new effective atomic number, Z e , which is based on published X-ray cross sections. Reference materials are used in conjunction with the system spectral response so that additional beam-hardening correction is not necessary. The technique is tested against other methods on DECT data of known specimens scanned by diverse spectra and systems. Uncertainties in accuracy and precision are less than 3% and 2% respectively for the (ρ e , Z e ) results compared to prior methods that are inaccurate and imprecise (over 9%).

42 ENGINEERING↗

Dynamic, resilient sensing system for automatic cyber-attack neutralization

An industrial asset may have monitoring nodes that generate current monitoring node values. An abnormality detection computer may determine that an abnormal monitoring node is currently being attacked or experiencing fault. A dynamic, resilient estimator constructs, using normal monitoring node values, a latent feature space (of lower dimensionality as compared to a temporal space) associated with latent features. The system also constructs, using normal monitoring node values, functions to project values into the latent feature space. Responsive to an indication that a node is currently being attacked or experiencing fault, the system may compute optimal values of the latent features to minimize a reconstruction error of the nodes not currently being attacked or experiencing a fault. The optimal values may then be projected back into the temporal space to provide estimated values and the current monitoring node values from the abnormal monitoring node are replaced with the estimated values.

97 MATHEMATICS AND COMPUTING↗

Dynamic concurrent learning method to neutralize cyber attacks and faults for industrial asset monitoring nodes

Input signals may be received from monitoring nodes of the industrial asset, each input signal comprising time series data representing current operation. A neutralization engine may transform the input signals into feature vectors in feature space, each feature vector being associated with one of a plurality of overlapping batches of received input signals. A dynamic decision boundary may be generated based on the set of feature vectors, and an abnormal state of the asset may be detected based on the set of feature vectors and a predetermined static decision boundary. An estimated neutralized value for each abnormal feature value may be calculated based on the dynamic decision boundary and the static decision boundary such that a future set of feature vectors will be moved with respect to the static decision boundary. An inverse transform of each estimated neutralized value may be performed to generate neutralized signals comprising time series data that are output.

97 MATHEMATICS AND COMPUTING↗

Machine Learning for Advanced Building Construction: Preprint

High-efficiency retrofits can play a key role in reducing carbon emissions associated with buildings if processes can be scaled-up to reduce cost, time, and disruption. Here we demonstrate an artificial intelligence/computer vision (AI/CV)- enabled framework for converting exterior build scans and dimensional data directly into manufacturing and installation specifications for overclad panels. In our workflow point clouds associated with LiDAR-scanned buildings are segmented into a facade feature space, vectorized features are extracted using an iterative random-sampling consensus algorithm, and from this representation an optimal panel design plan satisfying manufacturing constraints is generated. This system and the corresponding construction process is demonstrated on a test facade structure constructed at the National Renewable Energy Laboratory (NREL). We also include a brief summary of a techno-economic study designed to estimate the potential energy and cost impact of this new system.

building retrofits↗

Machine Learning for Advanced Building Construction

High-efficiency retrofits can play a key role in reducing carbon emissions associated with buildings if processes can be scaled-up to reduce cost, time, and disruption. Here we demonstrate an artificial intelligence/computer vision (AI/CV)-enabled framework for converting exterior build scans and dimensional data directly into manufacturing and installation specifications for overclad panels. In our workflow point clouds associated with LiDAR-scanned buildings are segmented into a facade feature space, vectorized features are extracted using an iterative random-sampling consensus algorithm, and from this representation an optimal panel design plan satisfying manufacturing constraints is generated. This system and the corresponding construction process is demonstrated on a test facade structure constructed at the National Renewable Energy Laboratory (NREL). We also include a brief summary of a techno-economic study designed to estimate the potential energy and cost impact of this new system.

build scans↗

A Data-driven Approach for Predicting Nepheline Crystallization in High-Level Waste Glasses

High-level waste (HLW) glasses with high aluminum content are prone to nepheline crystallization during the slow canister cooling that is experienced during large-scale production. Due to its detrimental effects on glass durability, nepheline precipitation must be avoided; however, developing robust, predictive models for nepheline crystallization behavior in HLW glasses is difficult due to their compositional complexity. Using overly conservative constraints to predict nepheline formation can limit the waste loading to lower than the achievable capacity. In this study, a robust data-driven model using five compositional features has been developed to predict nepheline formation. The analysis of the model and the data show that there is an overlap, instead of a distinct compositional boundary, between glasses that form and do not form nepheline. As a result, the model’s predictive accuracy is not the same throughout the feature space and instead is dependent on the location of the glass composition in the dimensionally reduced feature space.

Sargin, Irmak↗

Feature-adjacent multi-fidelity physics-informed machine learning for partial differential equations

Physics-informed neural networks have emerged as an alternative method for solving partial differential equations. However, for complex problems, the training of such networks can still require high-fidelity data which can be expensive to generate. To reduce or even eliminate the dependency on high-fidelity data, we propose a novel multi-fidelity architecture which is based on a feature space shared by the low- and high-fidelity solutions. In the feature space, the projections of the low-fidelity and high-fidelity solutions are adjacent by constraining their relative distance. The feature space is represented with an encoder and its mapping to the original solution space is effected through a decoder. In conclusion, the proposed multi-fidelity approach is validated on forward and inverse problems for steady and unsteady problems described by partial differential equations.

97 MATHEMATICS AND COMPUTING↗

Intelligent feature selection techniques for pattern classification of Lamb wave signals

Lamb wave interaction with flaws is a complex, three-dimensional phenomenon, which often frustrates signal interpretation schemes based on mode arrival time shifts predicted by dispersion curves. As the flaw severity increases, scattering and mode conversion effects will often dominate the time-domain signals, obscuring available information about flaws because multiple modes may arrive on top of each other. Even for idealized flaw geometries the scattering and mode conversion behavior of Lamb waves is very complex. Here, multi-mode Lamb waves in a metal plate are propagated across a rectangular flat-bottom hole in a sequence of pitch-catch measurements corresponding to the double crosshole tomography geometry. The flaw is sequentially deepened, with the Lamb wave measurements repeated at each flaw depth. Lamb wave tomography reconstructions are used to identify which waveforms have interacted with the flaw and thereby carry information about its depth. Multiple features are extracted from each of the Lamb wave signals using wavelets, which are then fed to statistical pattern classification algorithms that identify flaw severity. In order to achieve the highest classification accuracy, an optimal feature space is required but it’s never known a priori which features are going to be best. For structural health monitoring we make use of the fact that physical flaws, such as corrosion, will only increase over time. This allows us to identify feature vectors which are topologically well-behaved by requiring that sequential classes “line up” in feature vector space. An intelligent feature selection routine is illustrated that identifies favorable class distributions in multi-dimensional feature spaces using computational homology theory. Betti numbers and formal classification accuracies are calculated for each feature space subset to establish a correlation between the topology of the class distribution and the corresponding classification accuracy.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Harnessing structural stochasticity in the computational discovery and design of microstructures

This paper presents a deep generative model-based design methodology for tailoring the structural stochasticity of microstructures. Although numerous methods have been established for designing deterministic (periodic) or stochastic microstructures, a systematic design approach that allows the unified treatment of both deterministic and stochastic microstructure design domains has yet to be created. The proposed methodology resolves this issue by learning a unified feature space that embodies diverse structural patterns with continuously varying stochasticity levels. A highly diverse microstructure database is established to incorporate various types of deterministic and stochastic microstructure patterns. A property-aware deep generative model is proposed to learn a unified feature space of the structural characteristics, as well as the relationship between structure features and properties of interest. Autoencoder (AE), Variational Autoencoder (VAE), and Adversarial Autoencoder (AAE) are compared to understand their relative merits in the property-aware learning of the unified feature space. Microstructural designs with tailorable stochasticity and properties are obtained by searching the unified feature space. Multiple design cases are presented to demonstrate the capability of designing microstructures for structural stochasticity and properties. Furthermore, the proposed method is employed to create stochastically graded structures, which manipulate the mechanical behaviors by varying the local stochasticity of the structure.

36 MATERIALS SCIENCE↗

PySFD: comprehensive molecular insights from significant feature differences detected among many simulated ensembles

Motivation Many modeling analyses of molecular dynamics (MD) simulations are based on a definition of states that can be (groups of) clusters of simulation frames in a feature space composed of molecular coordinates. With increasing dimension of this feature space (due to the increasing size or complexity of a simulated molecule), it becomes very difficult to cluster the underlying MD data and estimate a statistically robust model. To mitigate this “curse of dimensionality”, one can reduce the feature space, e.g., with principal component or time-lagged independent component analysis transformations, focusing the analysis on the most important modes of transitions. In practice, however, all these reduction strategies may neglect important molecular details that are susceptible to experimental verification. Results To recover such molecular details, I have developed PySFD (Significant Feature Differences analyzer for Python), a multi-processing software package that efficiently selects significantly different features of any user-defined feature type among potentially many different simulated state ensembles, such as meta-stable states of a Markov State Model (MSM). Applying PySFD on MSMs of an aggregate of 300 microseconds MD simulations recently performed on the major histocompatibility complex class II (MHCII) protein, I demonstrate how this toolkit can extract and visualize valuable mechanistic information from big MD simulation data, e.g., in form of networks of dynamic interaction changes connecting functionally relevant sites of a protein complex.

Biochemistry & Molecular Biology↗