Search

We consider the problem of estimating the spatial layout of an indoor scene from a monocular RGB image, modeled as the projection of a 3D cuboid. Existing solutions to this problem often rely strongly on hand-engineered features and vanishing point detection, which are prone to failure in the presence of clutter. In this paper, we present a method that uses a fully convolutional neural network (FCNN) in conjunction with a novel optimization framework for generating layout estimates. We demonstrate that our method is robust in the presence of clutter and handles a wide range of highly challenging scenes. We evaluate our method on two standard benchmarks and show that it achieves state of the art results, outperforming previous methods by a wide margin.

Magnetoelectric behavior of a yttrium iron garnet (YIG)/zirconate titanate (PZT) magnetoelectric bilayer composite was studied over 1-7 GHz under different bias magnetic fields and electric fields by using a broadband air-gap microstrip with the PZT/YIG loaded in the air gap. Electrostatically induced ferrimagnetic resonance (FMR) frequency shifts of the YIG/PZT bilayer composite were studied. The FMR frequency shift was negligible at bias fields when the YIG was not saturated. After saturation, the FMR frequency increased nearly linearly from 15 MHz at a bias field of 100 Oe to 30 MHz at 1200 Oe and dropped suddenly at a field of 1300 Oe to about 20 MHz. This nonlinear bias magnetic held dependence was due to magnetic domain activities when the YIG was not saturated and due to the interference between the uniform mode and other magnetostatic spin waves after saturation. An electrostatically tunable band-reject filter device was demonstrated which has a peak attenuation of greater than 50 dB, 40 dB rejection band of 10 MHz, and pass band insertion loss of

This paper describes a capstone design project by four undergraduate students (first four authors listed in alphabetical order by last name). A noninvasive brain computer interface (BCI) based on the steady state visual evoked potential (SSVEP) has been developed and utilized in controlling an iRobot platform remotely in real-time closed-loop fashion using video feedback from the robot's eye view to the operator over the internet. The operator selects commands by focusing gaze on one of four flickering checkerboards surrounding the video feedback window in order to navigate the robot as desired. The intended/desired control commands are sent to a laptop controlling the iRobot platform via remote wireless connection. Naive subjects are able to control and navigate the robot via the designed interface with minimal practice and classifier calibration. Typical command selection accuracy is over 95% within 4 seconds of the desired transition; most subjects are able to achieve such high accuracies with a 1 second delay.

Ocean acoustic waveguide remote sensing (OAWRS) is a bistatic multibeam wide area sonar system that enables unaliased monitoring of fish populations over ecosystem scales. This allows us to quantify the formation processes of vast shoals of Atlantic herring (Clupea harengus) containing hundreds of millions of individuals, imaged during the Autumn 2006 spawning season. Areal population density estimation requires calibration of the low‐frequency target strength of an individual shoaling herring. This is estimated from experimental data acquired simultaneously at multiple frequencies in the 300–1200‐Hz range using (1) the OAWRS system, (2) areal population density calibration with several conventional fish finding sonar systems (CFFS), (3) fish length distributions obtained from trawl samples, and (4) local low‐frequency transmission loss measurements. High spatial‐temporal co‐registration was found between shoals imaged by OAWRS and concurrent CFFS line transects, which also provided fish depth distributio...

The mean low‐frequency target strength (TS) of spawning Atlantic herring populations in the Gulf of Maine is estimated from the experimental data acquired during September–October 2006 near the northern flank of Georges Bank. A low‐frequency OAWRS system with an instantaneous imaging diameter of 100 km was deployed to provide spatially unaliased imaging of fish populations over wide areas. The OAWRS system’s scattering strength measurements are calibrated with areal fish population density estimates obtained from concurrent localized line‐transect measurements with several conventional fish finding sonars (CFFSs). Trawl sampling at selected locations enables the identification of the imaged species. The mean TS estimates of herring individuals exhibits significant variation over OAWRS operating frequency range, in accordance with the results from a resonant scattering model for swimbladder‐bearing fish. The neutral buoyancy depth of herring and the species composition in the imaged population is inferred ...