Hasil untuk "Maps"

Ming-Fang Chang, John Lambert, Patsorn Sangkloy et al.

We present Argoverse, a dataset designed to support autonomous vehicle perception tasks including 3D tracking and motion forecasting. Argoverse includes sensor data collected by a fleet of autonomous vehicles in Pittsburgh and Miami as well as 3D tracking annotations, 300k extracted interesting vehicle trajectories, and rich semantic maps. The sensor data consists of 360 degree images from 7 cameras with overlapping fields of view, forward-facing stereo imagery, 3D point clouds from long range LiDAR, and 6-DOF pose. Our 290km of mapped lanes contain rich geometric and semantic metadata which are not currently available in any public dataset. All data is released under a Creative Commons license at Argoverse.org. In baseline experiments, we use map information such as lane direction, driveable area, and ground height to improve the accuracy of 3D object tracking. We use 3D object tracking to mine for more than 300k interesting vehicle trajectories to create a trajectory forecasting benchmark. Motion forecasting experiments ranging in complexity from classical methods (k-NN) to LSTMs demonstrate that using detailed vector maps with lane-level information substantially reduces prediction error. Our tracking and forecasting experiments represent only a superficial exploration of the potential of rich maps in robotic perception. We hope that Argoverse will enable the research community to explore these problems in greater depth.

K. Simonyan, A. Vedaldi, Andrew Zisserman

This paper addresses the visualisation of image classification models, learnt using deep Convolutional Networks (ConvNets). We consider two visualisation techniques, based on computing the gradient of the class score with respect to the input image. The first one generates an image, which maximises the class score [Erhan et al., 2009], thus visualising the notion of the class, captured by a ConvNet. The second technique computes a class saliency map, specific to a given image and class. We show that such maps can be employed for weakly supervised object segmentation using classification ConvNets. Finally, we establish the connection between the gradient-based ConvNet visualisation methods and deconvolutional networks [Zeiler et al., 2013].

M. Hansen, P. Potapov, R. Moore et al.

T. Mikkelsen, Manching Ku, David B. Jaffe et al.

M. Rosvall, Carl T. Bergstrom

To comprehend the multipartite organization of large-scale biological and social systems, we introduce an information theoretic approach that reveals community structure in weighted and directed networks. We use the probability flow of random walks on a network as a proxy for information flows in the real system and decompose the network into modules by compressing a description of the probability flow. The result is a map that both simplifies and highlights the regularities in the structure and their relationships. We illustrate the method by making a map of scientific communication as captured in the citation patterns of >6,000 journals. We discover a multicentric organization with fields that vary dramatically in size and degree of integration into the network of science. Along the backbone of the network—including physics, chemistry, molecular biology, and medicine—information flows bidirectionally, but the map reveals a directional pattern of citation from the applied fields to the basic sciences.

D. Schlegel, D. Finkbeiner, M. Davis

J. Barrett, B. Fry, J. Maller et al.

Karl J. Friston, A. Holmes, K. Worsley et al.

T. Jones, J. Zou, S. Cowan et al.

T. Kohonen

Eric S Lander, Eric S Lander, P. Green et al.

R. Voorrips

T. Meuwissen, B. Hayes, M. Goddard

T. Kohonen

S. Eickhoff, K. Stephan, H. Mohlberg et al.

E. Lander, D. Botstein

C. Bennett, M. Halpern, G. Hinshaw et al.

We present full-sky microwave maps in five frequency bands (23-94 GHz) from the Wilkinson Microwave Anisotropy Probe (WMAP) first-year sky survey. Calibration errors are less than 0.5%, and the low systematic error level is well specified. The cosmic microwave background (CMB) is separated from the foregrounds using multifrequency data. The sky maps are consistent with the 7° FWHM Cosmic Background Explorer (COBE) maps. We report more precise, but consistent, dipole and quadrupole values. The CMB anisotropy obeys Gaussian statistics with -58 < fNL < 134 (95% confidence level [CL]). The 2 ≤ ℓ ≤ 900 anisotropy power spectrum is cosmic-variance-limited for ℓ < 354, with a signal-to-noise ratio greater than 1 per mode to ℓ = 658. The temperature-polarization cross-power spectrum reveals both acoustic features and a large-angle correlation from reionization. The optical depth of reionization is τ = 0.17 ± 0.04, which implies a reionization epoch of tr = 180 Myr (95% CL) after the big bang at a redshift of zr = 20 (95% CL) for a range of ionization scenarios. This early reionization is incompatible with the presence of a significant warm dark matter density. A best-fit cosmological model to the CMB and other measures of large-scale structure works remarkably well with only a few parameters. The age of the best-fit universe is t0 = 13.7 ± 0.2 Gyr. Decoupling was tdec = 379 kyr after the big bang at a redshift of zdec = 1089 ± 1. The thickness of the decoupling surface was Δzdec = 195 ± 2. The matter density of the universe is Ωmh2 = 0.135, the baryon density is Ωbh2 = 0.0224 ± 0.0009, and the total mass-energy of the universe is Ωtot = 1.02 ± 0.02. It appears that there may be progressively less fluctuation power on smaller scales, from WMAP to fine-scale CMB measurements to galaxies and finally to the Lyα forest. This may be accounted for with a running spectral index of scalar fluctuations, fitted as ns = 0.93 ± 0.03 at wavenumber k0 = 0.05 Mpc-1 (ℓeff ≈ 700), with a slope of dns/d ln k = -0.031 in the best-fit model. (For WMAP data alone, ns = 0.99 ± 0.04.) This flat universe model is composed of 4.4% baryons, 22% dark matter, and 73% dark energy. The dark energy equation of state is limited to w < -0.78 (95% CL). Inflation theory is supported with ns ≈ 1, Ωtot ≈ 1, Gaussian random phases of the CMB anisotropy, and superhorizon fluctuations implied by the temperature-polarization anticorrelations at decoupling. An admixture of isocurvature modes does not improve the fit. The tensor-to-scalar ratio is r(k0 = 0.002 Mpc-1) < 0.90 (95% CL). The lack of CMB fluctuation power on the largest angular scales reported by COBE and confirmed by WMAP is intriguing. WMAP continues to operate, so results will improve.

C. Genovese, N. Lazar, Thomas E. Nichols

G. Abecasis, S. Cherny, W. Cookson et al.