Search NASA⌕ Search

SEARCH · Search NASA

Results for “data classification”

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 1,495 records · Page 83

An application of expert systems technology to remotely sensed image analysis

In October of 1983, a project was initiated at NASA/Ames Research Center to design a prototype expert system capable of producing a preliminary land cover classification from an unsupervised classification of Landsat Multispectral Scanner (MSS) imagery and associated collateral data. This paper outlines the problem of determining land cover classes from an unsupervised classification, presents a brief overview of expert systems, and continues with a description of the design and current state of implementation of NASA/Ames' prototype Landsat MSS image analysis expert system.

Erickson, W. K.↗

NASA's Meteoroid Environments Office's Response to Three Significant Bolide Events Over North America

Being the only U.S. Government entity charged with monitoring the meteor environment, the Meteoroid Environment Office has deployed a network of all sky and wide field meteor cameras, along with the appropriate software tools to quickly analyze data from these systems. However, the coverage of this network is still quite limited, forcing the incorporation of data from other cameras posted to the internet in analyzing many of the fireballs reported by the public and media. A procedure has been developed that determines the analysis process for a given fireball event based on the types and amount of data available. The differences between these analysis process will be explained and outlined by looking at three bolide events, all of which were large enough to produce meteorites. The first example is an ideal event - a bright meteor that occurred over NASA's All Sky Camera Network on August 2, 2014. With clear video of the event from various angles, a high-accuracy trajectory, beginning and end heights, orbit and approximate brightness/size of the event are able to be found very quickly using custom software. The bolide had the potential to have dropped meteorites, so dark flight analysis and modeling was performed, allowing potential fall locations to be mapped as a function of meteorite mass. The second case study was a bright bolide that occurred November 3, 2014 over West Virginia. This was just north of the NASA southeastern all-sky network, and just south of the Ohio-Pennsylvania network. This case study showcases the MEO's ability to use social media and various internet sources to locate videos of the event from obscure sources (including the Washington Monument) for anything that will permit a determination of a basic trajectory and fireball light curve The third case study will highlight the ability to use doppler weather radar in helping locate meteorites, which enable a definitive classification of the impactor. The input data and analysis steps differ for each case study, but the goals remain the same - a trajectory, orbit, and mass estimate for the bolide within hours of the event, and, for events with a high probability of producing meteorites, a location of the strewn field within a day.

Blaauw, Rhiannon C.↗

Spectral feature design in high dimensional multispectral data

The High resolution Imaging Spectrometer (HIRIS) is designed to acquire images simultaneously in 192 spectral bands in the 0.4 to 2.5 micrometers wavelength region. It will make possible the collection of essentially continuous reflectance spectra at a spectral resolution sufficient to extract significantly enhanced amounts of information from return signals as compared to existing systems. The advantages of such high dimensional data come at a cost of increased system and data complexity. For example, since the finer the spectral resolution, the higher the data rate, it becomes impractical to design the sensor to be operated continuously. It is essential to find new ways to preprocess the data which reduce the data rate while at the same time maintaining the information content of the high dimensional signal produced. Four spectral feature design techniques are developed from the Weighted Karhunen-Loeve Transforms: (1) non-overlapping band feature selection algorithm; (2) overlapping band feature selection algorithm; (3) Walsh function approach; and (4) infinite clipped optimal function approach. The infinite clipped optimal function approach is chosen since the features are easiest to find and their classification performance is the best. After the preprocessed data has been received at the ground station, canonical analysis is further used to find the best set of features under the criterion that maximal class separability is achieved. Both 100 dimensional vegetation data and 200 dimensional soil data were used to test the spectral feature design system. It was shown that the infinite clipped versions of the first 16 optimal features had excellent classification performance. The overall probability of correct classification is over 90 percent while providing for a reduced downlink data rate by a factor of 10.

Chen, Chih-Chien Thomas↗

Polarized Radar for Detection and Automatic Non-Visual Assessment of Unmanned Aerial Systems

This Laboratory Directed Research and Development (LDRD) effort performed fundamental Research and Development (R&D) to develop a robust radar processing algorithm capable of assessing the difference between an Unmanned Aerial System (UAS) and a biological target such as a bird, based on mathematics applied to the polarized radar returns of the target object, alone. The current threats of using such a UAS as a delivery platform for a host of destructive components is a major concern for the protection of various assets. Most recently, on 14th Sept. 2019, dozens of suicide or kamikaze drones (UAV-X) coordinated an attack on two Saudi oil facilities that demonstrated the potential to disrupt global oil supplies. While radar-based UAS detection systems can detect UAS at ranges greater than 1-km, the issues of excessive Nuisance/False Alarm Rates (NAR/FAR) from natural sources (birds in particular) has not been sufficiently addressed. In this effort we describe and utilize the Adaptive Polarization Difference Imaging-based (APDI) algorithms for the detection and automatic non-visual assessment of Unmanned Aerial System applications. Originally developed for optical imaging and sensing of polarization information in nature, the algorithms developed here are modified to serve for the target detection purposes in counter-UAS (cUAS) environments. We exploit the polarization statistics of the observing scene for detection and identification of changes within the scene and assess from these changes for UAS/bird classifications. Several cases are considered from independent data sources, including numerically generated data, anechoic chamber data as well as experimental radar data, to show the applicability of the techniques developed here. The methods developed in this effort are designed to be used in cUAS setups but have shown promise for a multitude of other radar-based classification uses as well.

47 OTHER INSTRUMENTATION↗

BOREAS AFM-12 1-km AVHRR Seasonal Land Cover Classification

The Boreal Ecosystem-Atmosphere Study (BOREAS) Airborne Fluxes and Meteorology (AFM)-12 team's efforts focused on regional scale Surface Vegetation and Atmosphere (SVAT) modeling to improve parameterization of the heterogeneous BOREAS landscape for use in larger scale Global Circulation Models (GCMs). This regional land cover data set was developed as part of a multitemporal one-kilometer Advanced Very High Resolution Radiometer (AVHRR) land cover analysis approach that was used as the basis for regional land cover mapping, fire disturbance-regeneration, and multiresolution land cover scaling studies in the boreal forest ecosystem of central Canada. This land cover classification was derived by using regional field observations from ground and low-level aircraft transits to analyze spectral-temporal clusters that were derived from an unsupervised cluster analysis of monthly Normalized Difference Vegetation Index (NDVI) image composites (April-September 1992). This regional data set was developed for use by BOREAS investigators, especially those involved in simulation modeling, remote sensing algorithm development, and aircraft flux studies. Based on regional field data verification, this multitemporal one-kilometer AVHRR land cover mapping approach was effective in characterizing the biome-level land cover structure, embedded spatially heterogeneous landscape patterns, and other types of key land cover information of interest to BOREAS modelers.The land cover mosaics in this classification include: (1) wet conifer mosaic (low, medium, and high tree stand density), (2) mixed coniferous-deciduous forest (80% coniferous, codominant, and 80% deciduous), (3) recent visible bum, vegetation regeneration, or rock outcrops-bare ground-sparsely vegetated slow regeneration bum (four classes), (4) open water and grassland marshes, and (5) general agricultural land use/ grasslands (three classes). This land cover mapping approach did not detect small subpixel-scale landscape features such as fens, bogs, and small water bodies. Field observations and comparisons with Landsat Thematic Mapper (TM) suggest a minimum effective resolution of these land cover classes in the range of three to four kilometers, in part, because of the daily to monthly compositing process. In general, potential accuracy limitations are mitigated by the use of conservative parameterization rules such as aggregation of predominant land cover classes within minimum horizontal grid cell sizes of ten kilometers. The AFM-12 one-kilometer AVHRR seasonal land cover classification data are available from the Earth Observing System Data and Information System (EOSDIS) Oak Ridge National Laboratory (ORNL) Distributed Active Archive Center (DAAC). The data files are available on a CD-ROM (see document number 20010000884).

Steyaert, Lou↗

Investigation of rainfall data with regard to low-level wind flow regime for east central Florida

Previous research has been conducted to investigate the effect of the low-level wind region on summertime convective storms in the east central Florida area. These effects were described by analyzing the distribution of lightning flashes within classifications based on the low-level wind regime for the months June through September of 1987 to 1990. The present research utilizes the same classification strategy to study rainfall patterns for data gathered for the CaPE (Convection and Precipitation/Electrification Experiment) field program. The CaPE field program was conducted in east central Florida from July 8, 1991 to August 18, 1991.

Brooks, Joni↗

NASA NeMO-Net

We present preliminary results from NASA NeMO-Net, the first neural multi-modal observation and training network for global coral reef assessment. NeMO-Net is an open-source deep convolutional neural network (CNN) and interactive active learning training software in development which will assess the present and past dynamics of coral reef ecosystems. NeMO-Net exploits active learning and data fusion of mm-scale remotely sensed 3D images of coral reefs captured using fluid lensing with the NASA FluidCam instrument, presently the highest-resolution remote sensing benthic imaging technology capable of removing ocean wave distortion, as well as hyperspectral airborne remote sensing data from the ongoing NASA CORAL mission and lower-resolution satellite data to determine coral reef ecosystem makeup globally at unprecedented spatial and temporal scales. Aquatic ecosystems, particularly coral reefs, remain quantitatively misrepresented by low- resolution remote sensing as a result of refractive distortion from ocean waves, optical attenuation, and remoteness. Machine learning classification of coral reefs using FluidCam mm-scale 3D data show that present satellite and airborne remote sensing techniques poorly characterize coral reef percent living cover, morphology type, and species breakdown at the mm, cm, and meter scales. Indeed, current global assessments of coral reef cover and morphology classification based on km-scale satellite data alone can suffer from segmentation errors greater than 40%, capable of change detection only on yearly temporal scales and decameter spatial scales, significantly hindering our understanding of patterns and processes in marine biodiversity at a time when these ecosystems are experiencing unprecedented anthropogenic pressures, ocean acidification, and sea surface temperature rise. NeMO-Net leverages our augmented machine learning algorithm that demonstrates data fusion of regional FluidCam (mm, cm-scale) airborne remote sensing with global low-resolution (m, km-scale) airborne and spaceborne imagery to reduce classification errors up to 80% over regional scales. Such technologies can substantially enhance our ability to assess coral reef ecosystems dynamics.

NASA↗

Comparison of Supervised and Un-Supervised Machine Learning Algorithms for Threat Detection and Scintillator Performance for Radiation Portal Monitoring

Following the events of September 11, 2001, international border crossing have been equipped with radiation portal monitors (RPMs) to identify illicit radioactive material. Polyvinyl toluene (PVT) scintillators are commonly used due to their low cost and reasonable maintainability, however they offer low spectral resolution. Despite the fact that over twenty years has transpired since this event, radioisotopes are still typically identified by hand-crafted classification algorithms, e.g., total counts or energy windowing, and exhibit relatively poor performance in detecting threats at the low false alarm rates required to support the stream of commerce. While some improvement to performance has been realized via the use of supervised machine learning, these classification algorithms typically utilize simulations in lieu of real data due to the sparsity of data for one or more classes. Accordingly, the performance of these algorithms is somewhat less than optimal when examining experiments or simulations with model mismatch. Consequently, in this work, we examine the application of a number of unsupervised machine learning, anomaly detection based algorithms, to circumvent the inverse crime when analyzing spectroscopy data for RPMs. We also compare anomaly detection results with those obtained via the use of supervised classification detection ML algorithms when model mismatch is introduced between the simulated threat items utilized for training/testing. Finally, we compared the performance of the PVT scintillators to those obtained with higher resolution detectors using both anomaly detection and supervised classification algorithms.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

A Preliminary Investigation of Systematic Noise in Data Acquired with the Airborne Imaging Spectrometer

Systematic noise is present in Airborne Imaging Spectrometer (AIS) data collected on October 26, 1983 and May 5, 1984 in grating position 0 (1.2 to 1.5 microns). In the October data set the noise occurs as 135 scan lines of low DN's every 270 scan lines. The noise is particularly bad in bands nine through thirty, restricting effective analysis to at best ten of the 32 bands. In the May data the regions of severe noise have been eliminated, but systematic noise is present with three frequencies (3, 106 and 200 scan lines) in all thirty two bands. The periodic nature of the noise in both data sets suggests that it could be removed as part of routine processing. This is necessary before classification routines or statistical analyses are used with these data.

Masuoka, E.↗

ExploreNEOs. V. AVERAGE ALBEDO BY TAXONOMIC COMPLEX IN THE NEAR-EARTH ASTEROID POPULATION

Examining the albedo distribution of the near-Earth object (NEO) population allows for a better understanding of the relationship between absolute (H) magnitude and size, which impacts calculations of the size frequency distribution and impact hazards. Examining NEO albedos also sheds light on the differences between the NEO and Main Belt populations. We combine albedo results from the ExploreNEOs Warm Spitzer Exploration Science program with taxonomic classifications from the literature, publicly available data sets, and new observations from our concurrent spectral survey to derive the average albedos for C-, D-, Q-, S-, V-, and X-complex NEOs. Using a sample size of 118 NEOs, we calculate average albedos of 0.29{sup +0.05}{sub -0.04}, 0.26{sup +0.04}{sub -0.03}, and 0.42{sup +0.13}{sub -0.11} for the Q-, S-, and V-complexes, respectively. The averages for the C- and D-complexes are 0.13{sup +0.06}{sub -0.05} and 0.02{sup +0.02}{sub -0.01}, but these averages are based on a small number of objects (five and two, respectively) and will improve with additional observations. We use albedos to assign X-complex asteroids to one of the E-, M-, or P-types. Our results demonstrate that the average albedos for the C-, S-, V-, and X-complexes are higher for NEOs than the corresponding averages observed in the Main Belt.

79 ASTRONOMY AND ASTROPHYSICS↗

Counterpart identification and classification for eRASS1 and characterisation of the active galactic nuclei content

Context. Accurately accounting for the Active Galactic Nucleus (AGN) phase in galaxy evolution requires a large, clean AGN sample. This is now possible with SRG/eROSITA, which completed its first all-sky X-ray survey (eRASS1) on June 12, 2020. The public Data Release 1 (DR1, Jan 31, 2024) includes 930,203 sources from the western Galactic hemisphere. Aims. The data enable the selection of a large AGN sample and the discovery of rare sources. However, scientific return depends on accurate characterisation of the X-ray emitters, requiring high-quality multi-wavelength data. This paper presents the identification and classification of optical and infrared counterparts to eRASS1 sources. Methods. Counterparts to eRASS1 X-ray point sources were identified using Gaia DR3, CatWISE2020, and Legacy Survey DR10 (LS10) with the Bayesian NWAY algorithm and trained priors. Sources were classified as Galactic or extragalactic via a machine-learning model combining optical/IR and X-ray properties, trained on a reference sample. For extragalactic LS10 sources, photometric redshifts were computed using CIRCLEZ. Results. Within the LS10 footprint, all 656,614 eROSITA/DR1 sources have at least one possible optical counterpart; ∼570 000 are extragalactic and likely AGN. Half are new detections compared to AllWISE, Gaia, and Quaia AGN catalogues. Gaia and CatWISE2020 counterparts are less reliable, due to the survey’s shallowness and the limited amount of features available to assess the probability of being an X-ray emitter. In the Galactic plane, where the overdensity of stellar sources also increases the chance of associations, using conservative reliability cuts, we identified approximately 18 000 Gaia and 55 000 CatWISE2020 extragalactic sources. Conclusions. We have released three high-quality counterpart catalogues – plus the training and validation sets – as a benchmark for the field. These datasets have many applications, but in particular, they empower researchers to build AGN samples tailored for completeness and purity, accelerating the hunt for the Universe’s most energetic engines.

X-rays: general↗

Arizona Water Resources: Utilizing Aerial Imagery and NASA Earth Observations to Assess Pinyon-Juniper Tree Mortality in Flagstaff, AZ

Pinyon-juniper woodlands (PJW) are a vital habitat and food source for several wildlife species and a source of both utility and cultural importance for Indigenous groups. In 2021, amidst a decades-long drought, an extensive juniper mortality event occurred at Wupatki National Monument (WNM) in Arizona. In response, the National Park Service (NPS) is evaluating which land management practices will be beneficial. In partnership with the NPS, the NASA DEVELOP team used remote sensing data to map PJW mortality and analyze the relation of tree mortality to stand density, climate, and topography in north-central Arizona from 2015 to 2021. To identify the extent of PJW, the team performed an unsupervised classification using National Agricultural Imagery Program (NAIP) data with validation sources including NPS-created land cover maps, Landscape Fire and Resource Management Planning Tools (LANDFIRE), NPS and United States Forest Service (USFS) vegetation maps, and Landsat 8 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) data. Terra Moderate Resolution Imaging Spectroradiometer (MODIS), Global Precipitation Measurement (GPM) Integrated Multi-satellite Retrievals for GPM (IMERG), Soil Moisture Active Passive (SMAP), Shuttle Radar Topography Mission (SRTM), and Landsat 8-derived Normalized Difference Vegetation Index (NDVI) and Normalized Difference Moisture Index (NDMI) were used to analyze factors contributing to pinyon-juniper mortality. Although no relationships were found in the broader study region, PJW mortality was weakly correlated to elevation, soil moisture, and land surface temperature within WNM. Results from this study can inform NPS vegetation management that best protects natural and cultural resources.

Margaret Jaenicke↗

Atmospheric Boundary Layer Classification With Doppler Lidar

We present a method using Doppler lidar data for identifying the main sources of turbulent mixing within the atmospheric boundary layer. The method identifies the presence of turbulence and then assigns a turbulent source by combining several lidar quantities: attenuated backscatter coefficient, vertical velocity skewness, dissipation rate of turbulent kinetic energy, and vector wind shear. Both buoyancy-driven and shear-driven situations are identified, and the method operates in both clear-sky and cloud-topped conditions, with some reservations in precipitation. To capture the full seasonal cycle, the classification method was applied to more than 1 year of data from two sites, Hyytiälä, Finland, and Jülich, Germany. Analysis showed seasonal variation in the diurnal cycle at both sites; a clear diurnal cycle was observed in spring, summer, and autumn seasons, but due to their respective latitudes, a weaker cycle in winter at Jülich, and almost non-existent at Hyytiälä. Additionally, there are significant contributions from sources other than convective mixing, with cloud-driven mixing being observed even within the first 500 m above ground. Also evident is the considerable amount of nocturnal mixing within the lowest 500 m at both sites, especially during the winter. The presence of a low-level jet was often detected when sources of nocturnal mixing were diagnosed as wind shear. The classification scheme and the climatology extracted from the classification provide insight into the processes responsible for mixing within the atmospheric boundary layer, how variable in space and time these can be, and how they vary with location.

54 ENVIRONMENTAL SCIENCES↗

Fast Query-Optimized Kernel-Machine Classification

A recently developed algorithm performs kernel-machine classification via incremental approximate nearest support vectors. The algorithm implements support-vector machines (SVMs) at speeds 10 to 100 times those attainable by use of conventional SVM algorithms. The algorithm offers potential benefits for classification of images, recognition of speech, recognition of handwriting, and diverse other applications in which there are requirements to discern patterns in large sets of data. SVMs constitute a subset of kernel machines (KMs), which have become popular as models for machine learning and, more specifically, for automated classification of input data on the basis of labeled training data. While similar in many ways to k-nearest-neighbors (k-NN) models and artificial neural networks (ANNs), SVMs tend to be more accurate. Using representations that scale only linearly in the numbers of training examples, while exploring nonlinear (kernelized) feature spaces that are exponentially larger than the original input dimensionality, KMs elegantly and practically overcome the classic curse of dimensionality. However, the price that one must pay for the power of KMs is that query-time complexity scales linearly with the number of training examples, making KMs often orders of magnitude more computationally expensive than are ANNs, decision trees, and other popular machine learning alternatives. The present algorithm treats an SVM classifier as a special form of a k-NN. The algorithm is based partly on an empirical observation that one can often achieve the same classification as that of an exact KM by using only small fraction of the nearest support vectors (SVs) of a query. The exact KM output is a weighted sum over the kernel values between the query and the SVs. In this algorithm, the KM output is approximated with a k-NN classifier, the output of which is a weighted sum only over the kernel values involving k selected SVs. Before query time, there are gathered statistics about how misleading the output of the k-NN model can be, relative to the outputs of the exact KM for a representative set of examples, for each possible k from 1 to the total number of SVs. From these statistics, there are derived upper and lower thresholds for each step k. These thresholds identify output levels for which the particular variant of the k-NN model already leans so strongly positively or negatively that a reversal in sign is unlikely, given the weaker SV neighbors still remaining. At query time, the partial output of each query is incrementally updated, stopping as soon as it exceeds the predetermined statistical thresholds of the current step. For an easy query, stopping can occur as early as step k = 1. For more difficult queries, stopping might not occur until nearly all SVs are touched. A key empirical observation is that this approach can tolerate very approximate nearest-neighbor orderings. In experiments, SVs and queries were projected to a subspace comprising the top few principal- component dimensions and neighbor orderings were computed in that subspace. This approach ensured that the overhead of the nearest-neighbor computations was insignificant, relative to that of the exact KM computation.

Mazzoni, Dominic↗

Target selection for the DESI Peculiar Velocity Survey

ABSTRACT We describe the target selection and characteristics of the DESI Peculiar Velocity Survey, the largest survey of peculiar velocities (PVs) using both the fundamental plane (FP) and the Tully–Fisher (TF) relationship planned to date. We detail how we identify suitable early-type galaxies (ETGs) for the FP and suitable late-type galaxies (LTGs) for the TF relation using the photometric data provided by the DESI Legacy Imaging Survey DR9. Subsequently, we provide targets for 373 533 ETGs and 118 637 LTGs within the Dark Energy Spectroscopic Instrument (DESI) 5-yr footprint. We validate these photometric selections using existing morphological classifications. Furthermore, we demonstrate using survey validation data that DESI is able to measure the spectroscopic properties to sufficient precision to obtain PVs for our targets. Based on realistic DESI fibre assignment simulations and spectroscopic success rates, we predict the final DESI PV Survey will obtain ∼133 000 FP-based and ∼53 000 TF-based PV measurements over an area of 14 000 deg2. We forecast the ability of using these data to measure the clustering of galaxy positions and PVs from the combined DESI PV and Bright Galaxy Surveys (BGS), which allows for cancellation of cosmic variance at low redshifts. With these forecasts, we anticipate a 4 per cent statistical measurement on the growth rate of structure at z < 0.15. This is over two times better than achievable with redshifts from the BGS alone. The combined DESI PV and BGS will enable the most precise tests to date of the time and scale dependence of large-scale structure growth at z < 0.15.

79 ASTRONOMY AND ASTROPHYSICS↗

Predictive analytics of selections of russet potatoes

We explore the application of machine learning algorithms specifically to enhance the selection process of Russet potato (Solanum tuberosum L.) clones in breeding trials by predicting their suitability for advancement. This study addresses the challenge of efficiently identifying high-yield, disease-resistant, and climate-resilient potato varieties that meet processing industry standards. Leveraging manually collected data from trials in the state of Oregon, we investigate the potential of a wide variety of state-of-the-art binary classification models. The dataset includes 1086 clones, with data on 38 attributes recorded for each clone, focusing on yield, size, appearance, and frying characteristics, with several control varieties planted consistently across four Oregon regions from 2013 to 2021. We conduct a comprehensive analysis of the dataset that includes preprocessing, feature engineering, and imputation to address missing values. We focus on several key metrics such as accuracy, F1-score, and Matthews correlation coefficient (MCC) for model evaluation. The top-performing models, namely a feedforward neural network classifier (Neural Net), a histogram-based gradient boosting classifier (HGBC), and a support vector machine classifier (SVM), demonstrate consistent and significant results. To further validate our findings, we conducted a simulation study using the aims, data-generating mechanisms, estimands, methods, and performance measures (ADEMP) framework, simulating different data-generating scenarios to assess model robustness and performance through true positive, true negative, false positive, and false negative distributions, area under the receiver operating characteristic curve (AUC-ROC) and MCC. The simulation results highlight that non-linear models like SVM and HGBC consistently show higher AUC-ROC and MCC than logistic regression, thus outperforming the traditional linear model across various distributions, and emphasizing the importance of model selection and tuning in agricultural trials. Variable selection further enhances model performance and identifies influential features in predicting trial outcomes. The findings emphasize the potential of machine learning in streamlining the selection process for potato varieties, offering benefits such as increased efficiency, substantial cost savings, and judicious resource utilization. Our study contributes insights into precision agriculture and showcases the relevance of advanced technologies for informed decision-making in breeding programs.

60 APPLIED LIFE SCIENCES↗

A support vector machine classifier reduces interscanner variation in the HRCT classification of regional disease pattern in diffuse lung disease: Comparison to a Bayesian classifier

Purpose: To investigate the effect of using different computed tomography (CT) scanners on the accuracy of high-resolution CT (HRCT) images in classifying regional disease patterns in patients with diffuse lung disease, support vector machine (SVM) and Bayesian classifiers were applied to multicenter data. Methods: Two experienced radiologists marked sets of 600 rectangular 20 Multiplication-Sign 20 pixel regions of interest (ROIs) on HRCT images obtained from two scanners (GE and Siemens), including 100 ROIs for each of local patterns of lungs-normal lung and five of regional pulmonary disease patterns (ground-glass opacity, reticular opacity, honeycombing, emphysema, and consolidation). Each ROI was assessed using 22 quantitative features belonging to one of the following descriptors: histogram, gradient, run-length, gray level co-occurrence matrix, low-attenuation area cluster, and top-hat transform. For automatic classification, a Bayesian classifier and a SVM classifier were compared under three different conditions. First, classification accuracies were estimated using data from each scanner. Next, data from the GE and Siemens scanners were used for training and testing, respectively, and vice versa. Finally, all ROI data were integrated regardless of the scanner type and were then trained and tested together. All experiments were performed based on forward feature selection and fivefold cross-validation with 20 repetitions. Results: For each scanner, better classification accuracies were achieved with the SVM classifier than the Bayesian classifier (92% and 82%, respectively, for the GE scanner; and 92% and 86%, respectively, for the Siemens scanner). The classification accuracies were 82%/72% for training with GE data and testing with Siemens data, and 79%/72% for the reverse. The use of training and test data obtained from the HRCT images of different scanners lowered the classification accuracy compared to the use of HRCT images from the same scanner. For integrated ROI data obtained from both scanners, the classification accuracies with the SVM and Bayesian classifiers were 92% and 77%, respectively. The selected features resulting from the classification process differed by scanner, with more features included for the classification of the integrated HRCT data than for the classification of the HRCT data from each scanner. For the integrated data, consisting of HRCT images of both scanners, the classification accuracy based on the SVM was statistically similar to the accuracy of the data obtained from each scanner. However, the classification accuracy of the integrated data using the Bayesian classifier was significantly lower than the classification accuracy of the ROI data of each scanner. Conclusions: The use of an integrated dataset along with a SVM classifier rather than a Bayesian classifier has benefits in terms of the classification accuracy of HRCT images acquired with more than one scanner. This finding is of relevance in studies involving large number of images, as is the case in a multicenter trial with different scanners.

60 APPLIED LIFE SCIENCES↗

The lower subsidiary diffuse plasma resonances and the classification of radio emissions below the plasma frequency

Using previusly published data and newly scaled ionograms from the Alouette 2 and ISIS 1 experiments, the diffuse ionospheric resonances Dn, stimulated by topside sounders Dn, are classified. The classification also includes the lower subsidiary resonances Dn(-) (n = 1, 2, 3, and 4). It is shown that the Dn(-) frequencies are related to f(Dn) and f(H) by the expression f(Dn)- = sq rt of (f(Dn)-squared - f(H)-squared).

Osherovich, Vladimir A.↗