An official website of the United States government
In this paper, we propose a convolutional neural network (CNN) based, scenario-dependent and sensor (mobile device) adaptable hierarchical classification framework. Our proposed framework is designed to automatically categorize face data captured under various challenging conditions, before the FR algorithms (pre-processing, feature extraction and matching) are used. First, a unique multi-sensor database (using Samsung S4 Zoom, Nokia 1020, iPhone 5S and Samsung S5 phones) is collected containing face images indoors, outdoors, with yaw angle from -90 to +90 and at two different distances, i.e. 1 and 10 meters. To cope with pose variations, face detection and pose estimation algorithms are used for classifying the facial images into a frontal or a non-frontal class. Next, our proposed framework is used where tri-level hierarchical classification is performed as follows: Level 1, face images are classified based on phone type; Level 2, face images are further classified into indoor and outdoor images; and finally, Level 3 face images are classified into a close (1m) and a far, low quality, (10m) distance categories respectively. Experimental results show that classification accuracy is scenario dependent, reaching from 95 to more than 98% accuracy for level 2 and from 90 to more than 99% for level 3 classification. A set of experiments is performed indicating that, the usage of data grouping before the face matching is performed, resulted in a significantly improved rank-1 identification rate when compared to the original (all vs. all) biometric system.
more »
« less
Use of mobile measurements to investigate frontal structures in Mississippi
Two cases of differently oriented frontal systems within Mississippi are investigated using data from a mobile vehicle-mounted observing system in addition to standard atmospheric data sources. Results highlight the capability of the mobile system to diagnose thermodynamic features at a wide range of spatial scales. Widely recognized frontal characteristics are noted in the data, together with some variations. Variations include a lack of strong relationship between frontal position and rainfall bands when examined at small scales. In one case a seemingly anomalous narrow band of significantly lower humidity was identified within about 20 km of the front. These results are indicative of the need for multi-scale data sources and for careful consideration of departures from classical models of phenomena for specific cases.
more »
« less
- Award ID(s):
- 1644888
- PAR ID:
- 10300266
- Date Published:
- Journal Name:
- Journal of the Mississippi Academy of Sciences
- Volume:
- 65
- Issue:
- 2
- ISSN:
- 0076-9436
- Page Range / eLocation ID:
- 170-182
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract In this study, we investigate the effects caused by interplanetary (IP) shock impact angles on the subsequent grounddB/dtvariations during substorms. IP shock impact angles have been revealed as a major factor controlling the subsequent geomagnetic activity, meaning that shocks with small inclinations with the Sun‐Earth line are more likely to trigger higher geomagnetic activity resulting from nearly symmetric magnetospheric compressions. Such field variations are linked to the generation of geomagnetically induced currents (GICs), which couple to artificial conductors on the ground leading to deleterious consequences. We use a sub‐set of a shock data base with 237 events observed in the solar wind at L1 upstream of the Earth, and large arrays of ground magnetometers at stations located in North America and Greenland. The spherical elementary current system methodology is applied to the geomagnetic field data, and field‐aligned‐like currents in the ionosphere are derived. Then, such currents are inverted back to the ground anddB/dtvariations are computed. Geographic maps are built with these field variations as a function of shock impact angles. The main findings of this investigation are: (a) typicaldB/dtvariations (5–10 nT/s) are caused by shocks with moderate inclinations; (b) the more frontal the shock impact, the more intense and the more spatially defined the ionospheric current amplitudes; and (c) nearly frontal shocks trigger more intensedB/dtvariations with larger equatorward latitudinal expansions. Therefore, the findings of this work provide new insights for GIC forecasting focusing on nearly frontal shock impacts on the magnetosphere.more » « less
-
Abstract In shallow water systems like coral reefs, bottom friction is an important term in the momentum balance. Parameterizations of bottom friction require a representation of canopy geometry, which can be conceptualized as an array of discrete obstacles or a continuous surface. Here, we assess the implications of using obstacle‐ and surface‐based representations to estimate geometric properties needed to parameterize drag. We collected high‐resolution reef topography data using a scanning multibeam sonar that resolved individual coral colonies within a set of 100‐m2reef patches primarily composed of moundingPoritescorals. The topography measurements yielded 1‐cm resolution continuous surfaces consisting of a single elevation value for each position in a regular horizontal grid. These surfaces were analyzed by (1) defining discrete obstacles and quantifying their properties (dimensions, shapes), and (2) computing properties of the elevation field (root mean square (rms) elevations, rms slopes, spectra). We then computed the roughness density (i.e., frontal area per unit plan area) using both analysis approaches. The obstacle and surface‐based estimates of roughness density did not agree, largely because small‐scale topographic variations contributed significantly to total frontal area. These results challenge the common conceptualization of shallow‐water canopies as obstacle arrays, which may not capture significant contributions of high‐wavenumber roughness to total frontal area. In contrast, the full range of roughness length scales present in natural reefs is captured by the continuous surface representation. Parameterizations of bottom friction over reef topography could potentially be improved by representing the contributions of all length scales to total frontal area and drag.more » « less
-
Abstract Marine-terminating glaciers on the Antarctic Peninsula (AP) have retreated, accelerated and thinned in response to climate change in recent decades. Ocean warming has been implicated as a trigger for these changes in glacier dynamics, yet little data exist near glacier termini to assess the role of ocean warming here. We use remotely-sensed iceberg melt rates seaward of two glaciers on the eastern and six glaciers on the western AP from 2013 to 2019 to explore connections between variations in ocean conditions and glacier frontal ablation. We find iceberg melt rates follow regional ocean temperature variations, with the highest melt rates (mean ≈ 10 cm d −1 ) at Cadman and Widdowson glaciers in the west and the lowest melt rates (mean ≈ 0.5 cm d −1 ) at Crane Glacier in the east. Near-coincident glacier frontal ablation rates from 2014 to 2018 vary from ~450 m a −1 at Edgeworth and Blanchard glaciers to ~3000 m a −1 at Seller Glacier, former Wordie Ice Shelf tributary. Variations in iceberg melt rates and glacier frontal ablation rates are significantly positively correlated around the AP (Spearman's ρ = 0.71, p -value = 0.003). We interpret this correlation as support for previous research suggesting submarine melting of glacier termini exerts control on glacier frontal dynamics around the AP.more » « less
-
Human mobility models typically produce mobility data to capture human mobility patterns individually or collectively based on real-world observations or assumptions, which are essential for many use cases in research and practice, e.g., mobile networking, autonomous driving, urban planning, and epidemic control. However, most existing mobility models suffer from practical issues like unknown accuracy and uncertain parameters in new use cases because they are normally designed and verified based on a particular use case (e.g., mobile phones, taxis, or mobile payments). This causes significant challenges for researchers when they try to select a representative human mobility model with appropriate parameters for new use cases. In this paper, we introduce a MObility VERification framework called MOVER to systematically measure the performance of a set of representative mobility models including both theoretical and empirical models based on a diverse set of use cases with various measures. Based on a taxonomy built upon spatial granularity and temporal continuity, we selected four representative mobility use cases (e.g., the vehicle tracking system, the camera-based system, the mobile payment system, and the cellular network system) to verify the generalizability of the state-of-the-art human mobility models. MOVER methodically characterizes the accuracy of five different mobility models in these four use cases based on a comprehensive set of mobility measures and provide two key lessons learned: (i) For the collective level measures, the finer spatial granularity of the user cases, the better generalization of the theoretical models; (ii) For the individual-level measures, the lower periodic temporal continuity of the user cases, the theoretical models typically generalize better than the empirical models. The verification results can help the research community to select appropriate mobility models and parameters in different use cases.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- The DOI is not currently available.
