dedup-isc-ft-v107-score
float64 0.32
1
| uid
stringlengths 32
32
| text
stringlengths 0
32.5k
| paper_id
stringlengths 1
14
| original_image_filename
stringlengths 5
222
|
|---|---|---|---|---|
0.440526 | 2c0b1ad5bc3249e39826189390ebb188 | CRM-CWMs: 0.033±0.010 | 2012.14228 | crm_slot6_k0_vis9_plot0_028080.jpg |
0.461728 | b665640b77a94ce0924cbeecaa405009 | CRM-CWMs: 0.023±0.013 | 2012.14228 | crm_slot6_k0_vis9_plot0_075737.jpg |
0.47334 | f5065a9087884ede9820f398b33487e0 | CRM-CWMs: 0.015±0.006 | 2012.14228 | crm_slot6_k5_vis9_plot0_024825.jpg |
0.412043 | f16e2c1bb5d64462a23aaee5332ec62c | CWMs: 0.048±0.017 | 2012.14228 | cwm_slot4_k0_vis9_plot0_029266.jpg |
0.440311 | f8743468da9e42acb0baff05f48194bc | CWMs: 0.038±0.015 | 2012.14228 | cwm_slot4_k1_vis9_plot0_042923.jpg |
0.4844 | b4336620ef6a47a49f25a35f9fdfeb90 | CWMs: 0.027±0.006 | 2012.14228 | cwm_slot6_k1_vis9_plot0_029528.jpg |
0.438139 | b28f8f14c894467983260b42ad61e0a5 | CWMs: 0.049±0.016 | 2012.14228 | cwm_slot6_k2_vis9_plot0_048960.jpg |
0.43782 | c1789afd5d8c422fb68d811ffc279940 | CWMs: 0.038±0.013 | 2012.14228 | cwm_slot6_k2_vis9_plot0_053534.jpg |
0.4374 | 5b7bef568b094225821754ddee5e0d2e | 2012.14228 | graphical_model_confounder_new.jpg |
|
0.4374 | 898a58ac3e944e7cb79ae9dd15697b78 | 2012.14228 | graphical_model_do_new.jpg |
|
0.4374 | 8d41fbf5eab8400baead9774a266c723 | 2012.14228 | graphical_model_no_confounder_new.jpg |
|
0.506939 | c421e7dc509c41a1a9d1481a19a75b3f | (H@1) Ranking results for multi-step prediction in latent space in different environments. Our models (CWM and CRM-CWM) consistently achieve the best result. | 2012.14228 | h1.jpg |
0.455433 | 8c185ccd5c3c4c8d86a4f94820dcb0f1 | An illustration of the proposed Causal World Models. | 2012.14228 | model.jpg |
0.484917 | 1a8dd20493284641babc87a3c1fcc8e9 | (MRR) Ranking results for multi-step prediction in latent space in different environments. Our models (CWM and CRM-CWM) consistently achieve the best result. | 2012.14228 | mrr.jpg |
0.439426 | 34050c2b0c084ff1a2231290eab1e1d3 | 2012.14228 | plot.jpg |
|
0.462272 | accfef9b2018406fbe366d067e0ad261 | Example from <cit.>. | 2012.14228 | simpsons.jpg |
0.457814 | e5e77e3047024286ac9065f6baef00b5 | WMs: 0.058±0.015 | 2012.14228 | wm_slot2_k1_vis9_plot0_112059.jpg |
0.51174 | 7471059d80384a69bfef154bb8631e2e | WMs: 0.053±0.021 | 2012.14228 | wm_slot4_k0_vis9_plot0_045488.jpg |
0.491415 | 35b90b03da4e4b5fb83a926cec3c93fb | WMs: 0.053±0.017 | 2012.14228 | wm_slot4_k0_vis9_plot0_069490.jpg |
0.515094 | 63bb12db7d264147baa22e6e668e15ae | WMs: 0.096±0.030 | 2012.14228 | wm_slot6_k1_vis9_plot0_081310.jpg |
0.465507 | 619f16e28cbf4df897fdeb88c5fa3aca | WMs: 0.056±0.037 | 2012.14228 | wm_slot6_k4_vis9_plot0_105778.jpg |
0.422862 | 4ea2265a1a064bd2a509f15573a93317 | Sample characterization. a) Schematic image of the sample. The white areas are Co2MnGa. V_xx (AMR) are the contacts used for measusing voltage in the AMR experiment, while the current I was applied between V_xx (AMTP). In the AMTP experiment, the voltage was measured at V_xx (AMTP). The red colored areas are Pt wires which are used in the AMTP experiments as on-chip heater and thermometer (green). b) in-plane and out-of-plane measurement of the magnetization vs. applied magnetic field. The inset show the magnetization vs. temperature. The saturated magnetic moment is approx. 4 μ_B per formula unit. c) X-ray diffraction radial scans of Co2MnGa films with 40 and 50nm thickness. Bragg peaks labeled by an asterisk (*) originate from the (00L) series of the MgO substrate, while Bragg peaks of Co2MnGa are labeled by their Miller indices. d) X-ray reflectivity pattern of the 40 nm and 50 nm thick Co2MnGa films. Experimental data are shown as data points, while the solid line represents a model fit based on the extended Parratt formalism. | 2012.14229 | p01.jpg |
0.51436 | 907fcbf5be614f46a81714bd8fc1c12b | Results of the AMR and AMTP measurement (crosses) and fitting (continuous lines) in the 50 nm sample at 300 K (AMR) and 250 K (AMTP). a) b), c) Schematic sketch of the magnetic field rotation in the XY, ZY, ZX plane with respect to the coordinate system and the sample. d), e), f) AMR (blue, left y-axes) and AMTP (orange, right y-axes) results; note the shifts of vertical axes indicated at the top of each panel: minimum resistivity, which occurs for m∥j, is 171 μΩ.cm. | 2012.14229 | p02.jpg |
0.497766 | 4c2a6dcd094141de8eb56c61e0b65efb | Temperature evolution of the parameters obtained by the phenomenological fit to AMR data. (A) Uniaxial magnetic anisotropy, (B-F) parameters for AMR defined in Eq. <ref>. | 2012.14229 | p03.jpg |
0.489188 | e2bfdfa631664fa6942535968843f200 | (a,c,e) Data from Fig. <ref> where AMR (AMTP) is fitted by Eq. <ref> (Eq. <ref>, respectively) with second-order terms only; a lower quality, compared to Fig. <ref>(f), of the AMR fit in the bottom panel is clearly apparent but systematic deviations can also be observed for the XY and ZY rotations. (b,d,f) The difference of the experimental data and fits on the left (fit residuals), see text for a detailed discussion. | 2012.14229 | p04.jpg |
0.444791 | a75051f83100459f85a0022e90bb2711 | a) Longitudinal Seebeck coefficient Σ_0 as a function of the temperature. The different samples are distinguished by the shape of symbols. The absolute value of the Seebeck coefficient of the 50 nm sample is monotonously increasing with temperature. There is a sign change at low temperatures. The Seebeck coefficients at T = 300 K for the 40 and 50 nm sample are of comparable magnitude. b) Comparison of the AMTP and the corresponding AMR parameters in the 50 nm sample as a function of temperature. AMR and AMTP are distinguished by linestyle. The m_y^2 contributions are in shades of orange, the m_z^2 contributions in shades of green. They do not seem to follow any common trend. Error bars of AMTP and AMR parameters in b) (implied by the fitting procedure) are too small to be visually resolved. | 2012.14229 | p05.jpg |
0.493686 | b0667e9bea024a51b92a503e55d7024d | Registration accuracies of the Classical, CNN, and Segis-Net pipeline as quantified by spatial correlation (SC) of the registered tract density maps. Error bars indicate standard deviations. The bracket hat indicates a significant difference between two methods (t-test, p<2.8 × 10^-3). | 2012.14230 | CS_3pipelines_bar.jpg |
0.390386 | 5911f8d76ce64d9e85c20194afaade4a | The network architectures of 𝒢_Ψ module used in the present study. The target image (I_t) and affine-aligned source image (I_s ∘ϕ_A) are used as the input to predict non-rigid deformation (ϕ_D), which can subsequently lead to a composite displacement field (ϕ) as shown in Figure <ref>. The number of convolution kernels used in Conv Unit 01 is denoted as [K1, K2], and that for Conv Unit 03 as [K1]. | 2012.14230 | Reg-Net_legend.jpg |
0.40972 | ee8da1b3c02d4baf987c7d49dcd6d137 | The network architectures of ℱ_Θ module used in the present study. The number of convolution kernels used in Conv Unit 01 is denoted as [K1, ..., Kn], and that for Conv Unit 02 as [K1] | 2012.14230 | Seg-Net_legend.jpg |
0.999988 | 2339756e3ae04bc9a7ef82650819ff6c | A flowchart for the inclusion, exclusion, and split of the datasets. | 2012.14230 | flowchart_dataInclusion_box.jpg |
0.452119 | 4181f6c35a684e2ab2d0842f006ae46a | Overview of the Segis-Net framework. Θ and Ψ denote the parameters of the segmentation (ℱ_Θ) and registration (𝒢_Ψ) function, respectively. Black circle indicates spatial warp with affine matrix (ϕ_A) or the composite displacement field (ϕ̂). The concatenation of the affine-aligned images is used as the input for 𝒢_Ψ. Loss function consists of ℒ_seg, ℒ_com, ℒ_reg and ℒ_def terms. Solid lines indicate the primary workflow of the method; dashed lines indicate the operations that are only implemented during training and could be adapted for applications. | 2012.14230 | overview_dec19_reviseNotation.jpg |
0.444728 | 25b82fb348a84f46bb112640f3187783 | Segmentation accuracies of the CNN pipeline and Segis-Net for different tracts. Error bars indicate standard deviations. | 2012.14230 | segACC_3pipelines_bar.jpg |
0.540066 | ebb84f6cb62e4a4fb400f185d4b8613b | Spatio-temporal consistency of segmentation (STCS) with the Classical, CNN, and Segis-Net pipeline. Error bars indicate standard deviations. The bracket hat indicates a significant difference between two methods (t-test, p<2.8 × 10^-3). | 2012.14230 | selfConsistency_3pipelines_bar.jpg |
0.407447 | 7f988c74e66946a896351ce7135992c6 | (a) Pictorial representation of radiation torque shot noise. An anisotropic scatterer in a linearly polarized light field experiences a fluctuating torque which arises from the vacuum fluctuations entering the unused port of the polarizing beamsplitter (PBS). (b) Schematic of the experimental setup. Inside a vacuum chamber, we focus a laser beam (propagating along z, linearly polarized along x) with an aspheric lens (0.7NA) to form an optical trap. In the forward direction, the light is collected and split into two paths with a beamsplitter (BS). One half of the optical power is sent to a center-of-mass (COM) motion detector. The other half is used to measure the libration angle θ in a balanced detection scheme. The measurement is recorded with a data acquisition device (DAQ). The intensity of the laser beam [wavelength λ = 1565.0(1)] is modulated with an electro-optic modulator (EOM) using feedback signals derived from the COM and the libration detector, respectively. (c) The blue line shows the measured power spectral density S_θθ of the libration motion at a pressure of p_gas = 7.0(7). The broad spectrum is a result of coupling between the angular degrees of freedom. The black line shows S_θθ at p_gas = 1.1(1)E-8 and with feedback-cooling engaged for COM and librational motion, where the signal of the libration detector reduces to a single resonant line. | 2012.14231 | Libration_Setup_Paper_2_different_label.jpg |
0.459931 | 41054cafe172411fa39ad53acea69c64 | Reheating experiment at p_gas = 1.1(1)E-8. (a) Cooled libration spectrum right after feedback cooling is turned off. (b) Libration spectrum after 950, just before the feedback cooling is turned back on. (c) Libration energy (blue circles) as a function of time. A linear fit to the data is shown as the solid line. | 2012.14231 | Libration_spectrum_energy_time_paper_different_label.jpg |
0.480886 | 1d35331114c1463c8b3d6e293cbde0a0 | Heating rate as a function of RIN at 1.1(1)E-8 (blue circles). An effect on the heating rate is observable only for RIN values exceeding -125c . | 2012.14231 | RIN_added_heating_rate_paper.jpg |
0.447608 | a9e9db3f7b7e47438d47398d0dde5d52 | Heating rate (blue circles) as a function of pressure. The solid black curve shows a linear fit with constant offset Γ_res (dashed line). The dotted line indicates the contribution of the gas to the heating rate. The red line shows the theoretical prediction for the radiation torque shot noise heating rate Γ_sn. | 2012.14231 | Reheating_energy_pressure_paper.jpg |
0.462878 | 7c30e123f93a453c9561a752dcb96d94 | Overview of the auto weak supervision model. | 2012.14234 | overview.jpg |
0.463304 | de2f0654379349f2bc896273320e545c | Appearance rendering module. This module interprets rendered feature as RGB pixel values. M_f denotes the number of feature channels and R is the number of channels of the intermediate appearance features (R=70+M_f). The intermediate appearance features are refined by 5 convolutional blocks and decoded by a Feature Decoder. The Feature Decoder is implemented as a three-layer perceptron network with R/2, R/4, and 3 neurons respectively, each layer is followed by LeakyReLU activation function, except for the very last one that uses HardTanh to produce normalized RGB color values. | 2012.14240 | deepdurfels_appearance_rendering_supp.jpg |
0.460483 | d07d680cd1a64acdb9a7474d6d2a1698 | Fusion network architecture. This module is a part of our learned appearance fusion pipeline ( <ref>). It creates a blended feature map F̂_t that needs to be integrated into DeepSurfel representation.. | 2012.14240 | deepdurfels_fusion_network_supp.jpg |
0.395199 | ca89626490ae4aa494cf1d4c53af7077 | Novel-view synthesis on unseen real-world data <cit.>. Our DeepSurfel method with 4×4 patches is trained on the Lion scene for 80 training iterations and then evaluated on the Gate and the Bricks scenes. The images show novel viewpoints. | 2012.14240 | figure_vc.jpg |
0.454195 | 74adfac64930429ca3cd614f0cef70c8 | Overview of our learned appearance fusion pipeline. The pipeline consists of an Appearance Fusion module that integrates a new RGB measurement I_t into DeepSurfels S_t-1 and a differentiable Appearance Rendering module that interprets and renders the content of representation for a given viewpoint. White blocks denote differentiable deterministic operations, rectangular blocks denote data, rounded rectangular blocks are trainable modules, and ⊗ is a feature stacking operation. | 2012.14240 | full_pipeline2.jpg |
0.437655 | 443116f873b74bd28ca742b30dd9ebf1 | Qualitative results of our model on unseen scenes from ShapeNet <cit.>. Compared to deterministic rendering from RGB values, DeepSurfels with learned (3+3)-channel 6×6 patches on a sparse 32^3 grid yields significantly more high-frequency details. As it is shown in the close ups, storing learned features in the texels is particularly useful to correct discretization artefacts. | 2012.14240 | generalization2.jpg |
0.392406 | 6f889c6300a44694af3ec1fc27621078 | Novel view synthesis for Replica <cit.> indoor scenes. The figure shows different views on two scenes (left and right). Our learned approach has been trained only on the room on the left. NeRF <cit.> is optimized separately on both scenes. | 2012.14240 | indoor_new.jpg |
0.431708 | f4a0e8233bb340ba88d8e4055e3bbc7b | DeepSurfel surface fitting. In a recursive fitting procedure, we align the texels of each patch with the underlying SDF surface by shifting its location and adjusting its normal. d denotes queried signed distance, n denotes SDF gradient ∇SDF in x, y, z directions. | 2012.14240 | surfel_alignment.jpg |
0.447393 | afed5ce7881a4cbe9077db5ec51ceb9d | Overview of our online appearance fusion pipeline and the DeepSurfel scene representation. The Appearance Fusion network efficiently aggregates appearance information from a stream of camera views into the proposed DeepSurfel representation S_t-1 that maintains high-frequency geometric and appearance information. DeepSurfels is a sparse grid of 2D patches that consist of surface-aligned texels, which encode appearance information either as RGB color values or learned feature vectors. The proposed Appearance Rendering network interprets aggregated and interpolated geometric and appearance information stored in DeepSurfels for rendering novel viewpoints. In this example we used DeepSurfels with a sparse 64^3 patch grid with 8 × 8 resolution surfel patches. | 2012.14240 | teaser_sketch2.jpg |
0.476057 | 9739cc36c9404c7981bf780d6482bf73 | Schematic representation of the experimental setup. A master oscillator power amplifier (MOPA) setup, comprising an optical parametric oscillator (OPO) and an optical parametric amplifier (OPA), is pumped by a Nd:YAG laser (blue line). The signal beam is separated via polarization optics and the idler beam (λ=2.17 µm) is focused onto tin microdroplets within a vacuum chamber. EUV emission is captured by a transmission grating spectrometer positioned at 60 with respect to the laser axis. An additional, third KTP crystal (dashed outline) was used in the OPA in a subset of the experiments. | 2012.14243 | MOPA_layout_essentials_arisholm_layout_solid_target_v2.jpg |
0.445482 | bd07932904534319af52af4170c645b9 | Top: maximum temperature of a tin plasma for various laser intensities calculated with the two-dimensional-radiation transport code RALEF-2D. A 30-µm diameter droplet is illuminated with temporally and spatially Gaussian-shaped laser pulses of wavelengths 1 and 2 µm. Center: temperature and electron density lineout along the axis of the incoming laser beam. Bottom: frequency-integrated local radiation field intensity I_rad of the plasma and its normalized derivative dI_rad. The radiation field intensity is calculated from Eq (<ref>) using the density and temperature lineouts depicted in the center panel. For more detail see text. | 2012.14243 | RALEF_maxTemp.jpg |
0.596649 | 3b30c337b5db4d5cbd06b100cdaa1565 | Spectral emission from tin plasmas produced with 1- and 2-µm laser wavelength for small and large droplet diameters at laser intensities of 2.4 and 1.1e11, respectively. | 2012.14243 | droplet_size_MOPA_YAG.jpg |
0.498093 | c3c6aacfbd2a4d76bba09a4c3079d36d | Spectra from tin droplet plasma observed at various intensities of the (a) 2-µm beam (4.7 ns pulse duration, 106 µm FWHM, 3 KTP crystals) and the (b) 1-µm Nd:YAG beam (10 ns pulse duration, 86 µm FWHM). The droplet size in both cases is 30 µm. (c) Ratio of the intensities of 1- and 2-µm drive laser beams needed to obtain spectra with similar spectral features. The data points represent the average intensity ratios from data taken with four different laser spot sizes of the 2 µm laser beam of 65x88, 106, 152 and 194 µm (FWHM), respectively. The error bars indicate the standard deviation per measurement. The red line represents the average over all data points and the shaded band the standard deviation of the average. | 2012.14243 | intensity_scan.jpg |
0.499623 | 2e5aed50480d4b53bb3bcf24cbed2abc | (a) Dependency of peak optical depth τ_i,p on droplet diameter for 5- and 10-ns laser pulse duration at 1-µm wavelength and for 4.3-ns pulse duration at 2-µm wavelength. Circles indicate Gaussian spatial laser profile and boxes indicate a homogeneous 'flattop' laser beam profile. Peak optical depth is fitted with respect to the spectrum obtained at 1-µm wavelength, 10-ns pulse duration and 30-µm droplet diameter with optical depth of τ_0,p=4.5. (b) Experimental values for spectral purity (SP) versus peak optical depth. The dashed line represents SP as calculated from the radiation-transported reference spectrum. The diamond symbol indicates the SP value of the radiation-transported reference spectrum for a peak optical depth value τ_i,p=0.4, obtained from comparison of the reference spectrum with the emission of the CO_2-laser-driven plasma spectrum illustrated in Fig. <ref>. | 2012.14243 | optical_depth_QR_base_with_SNS201811_v2.jpg |
0.496586 | ec9b713910fe4e9cba3aca43b0694448 | Spectrum produced with 2-µm laser light (red line) compared to the radiation-transported reference spectrum for a peak optical depth value of τ_p=2.2 (gray line, barely distinguishable from the red line). Reference and 2-µm driven spectra were both obtained using a droplet diameter of 30 µm. Also shown is a spectrum obtained using a 10-µm CO_2 laser <cit.> that represents the case of small optical depth. | 2012.14243 | radiation_transport_scan4_30um_droplet.jpg |
0.4558 | d88bd5dd49674d0f9d46006275314219 | Cosmological observables n_s and r obtained by natural inflation with and without the nonminimal coupling. This figure shows the case where n=2 and 50-60 e-folds are achieved during the inflation. | 2012.14248 | Pn2xifixed2.jpg |
0.483239 | 19104f9339024d4d96ae0ef4ad2260fb | Behavior of the equation of state parameter ω as a function of the number of e-folds, using three different values of the coupling ξ. On the horizontal axis, the 0 corresponds to the end of inflation, negative values correspond to the inflationary epoch, and the positive values indicate post-inflationary epoch. In the plot we use n=2, f=3M_pl, in both the metric (solid line) and the Palatini (dashed lines) cases, their differences are very small. The results from other values of n and f are similar. | 2012.14248 | Preheat2.jpg |
0.43913 | 7ad6fb1ac84c42e19fcc0e6726e24fb1 | Plot of field χ as a function of ϕ with n=2. Color lines correspond to different values of the coupling ξ. | 2012.14248 | chiphi.jpg |
0.430246 | ba75e98a37484c8ea8e1817f2fe2d4f4 | Relations between the inflaton fields χ and ϕ with n=2 (solid lines) and n=4 (dashed lines) in the context of Palatini theory. Each color represents fixed value of ξ indicated inside the plot. | 2012.14248 | f1n24f2.jpg |
0.461245 | 6d3ebba078e94babaac67887d97a8e76 | Illustration of (a) the original ASIS module and (b) the Boundary-NET module. | 2012.14253 | ASIS_net.jpg |
0.391138 | 651c1026269e43da970195b87fa841e7 | (a) Ground truth and predicted instances with BFS for (b) ϵ=7cm and (c) ϵ=2cm. Results shown for the oil refinery. Each color in the figure represents one point cluster instance. For larger ϵ, we observe under-segmentation of instances, for smaller ϵ over-segmentation | 2012.14253 | BFS_7cm.jpg |
0.436573 | d021291a7bdd4db59ceb9c10c9a8c3a4 | (i) Ground truth annotated instances and (ii) predicted instances with BFS alogirthm for representative windows from the oil refinery.Each color in the figure represents one point cluster instance. | 2012.14253 | BFS_BP_gt.jpg |
0.432167 | d1525796ba194fdc87aa602e8b2b0225 | (a) BFS instance precision and (b) recall per CLOI class and (c) mean precision and recall for different IoU thresholds for the oil refinery. | 2012.14253 | BFS_BP_gt_class.jpg |
0.42171 | 295bafe0f7604b16b0b2a77a8557a7e4 | (i) Ground truth annotated instances and (ii) predicted instances with BFS alogirthm for representative windows from the petrochemical plant.Each color in the figure represents one point cluster instance. | 2012.14253 | BFS_EATON_gt.jpg |
0.43671 | 4335b01207d84fcbb802727b9382a28f | (a) BFS instance precision and (b) recall per CLOI class and (c) mean precision and recall for different IoU thresholds for the petrochemical plant. | 2012.14253 | BFS_EATON_gt_class.jpg |
0.42582 | bfa89e50be3b4374bb960356b8a1600b | (i) Ground truth annotated instances and (ii) predicted instances with BFS algorithm for representative windows from the processing unit.Each color in the figure represents one point cluster instance. | 2012.14253 | BFS_NF_gt.jpg |
0.397433 | 1311631cb1e34d12be5f31758427a8ad | (a) BFS instance precision and (b) recall per CLOI class and (c) mean precision and recall for different IoU thresholds for the processing unit. | 2012.14253 | BFS_NF_gt_class.jpg |
0.465866 | 44fafb68793d49678cf9275584c2dfee | Average precision (%) of the BFS algorithm with respect to the radius (ϵ) for each CLOI facility. | 2012.14253 | BFS_Prec_radius.jpg |
0.411229 | f45d30362a524182b60342c8aee35b71 | Average recall (%) of the BFS algorithm with respect to the radius (ϵ) for each CLOI facility. | 2012.14253 | BFS_Rec_radius.jpg |
0.407476 | 84a92948377d43648884f2a4203996d2 | (i) Ground truth annotated instances and (ii) predicted instances with BFS alogirthm for representative windows from the warehouse.Each color in the figure represents one point cluster instance. | 2012.14253 | BFS_THOR_gt.jpg |
0.447384 | 934aaf47d37c4458849d2eac2c05b636 | (a) BFS instance precision and (b) recall per CLOI class and (c) mean precision and recall for different IoU thresholds for the warehouse. | 2012.14253 | BFS_THOR_gt_class.jpg |
0.387973 | 2b2b01b7bc004b1ba09991d57714a47d | Boundary limitations in case of (a) conduit proximity and (b) noisy TLS data. Each color in the figure represents one point cluster instance. | 2012.14253 | BFS_noiseProximity_gt.jpg |
0.455964 | 2fc2e877c69348e496fefde6bfd80adf | BFS radius with respect to mRec_ins (%) for different IoU thresholds for each CLOI facility. | 2012.14253 | BFS_radius.jpg |
0.43409 | 8720496fe92b494aa0368b7c6ac97d2d | BFS radius with Rec (%) per CLOI class for IoU=50% for each CLOI facility. | 2012.14253 | BFS_radius_shapes.jpg |
0.4188 | 80bac3f7466f4b728428ad6f4b5bd2ab | Proposed instance segmentation methodology | 2012.14253 | Methodology.jpg |
0.478417 | a798cf5379b44865b4b95eefc3594581 | Performance of the BFS algorithm with respect to the minimum instance size (μ) for IoU=50% and ϵ=4 cm. Test on the oil refinery facility. | 2012.14253 | minInsSize.jpg |
0.430115 | a9663c4200464d3f8299a93a06d10c8c | (i) Ground truth annotated instances and (ii) predicted instances with BFS alogirthm for representative examples of valves and flanges across the CLOI facilities. Each color in the figure represents one point cluster instance. | 2012.14253 | valveflangeGT.jpg |
0.46203 | f43e5f3476ed4786a929658ddecc5176 | Uniform prior | 2012.14264 | scenario_0.jpg |
0.396507 | 86365cbb2b1a4ba2846dd4d5e73427a9 | Uniform prior | 2012.14264 | scenario_0.jpg |
0.378317 | b9411bf467904808b437b16e7a45c0bc | Uniform prior | 2012.14264 | scenario_0.jpg |
0.41688 | 19e550221c434b21a47e487860fbea93 | K=250 | 2012.14264 | scenario_0_K_250.jpg |
0.474814 | 99b2654286f642528c709bc70cfbe441 | K=250 | 2012.14264 | scenario_0_K_250.jpg |
0.399899 | 7ea0360c2455410d8635515d773aa6d9 | K=5 | 2012.14264 | scenario_0_K_5.jpg |
0.407616 | 7967d886e3404d1a9c73a767aba36219 | K=5 | 2012.14264 | scenario_0_K_5.jpg |
0.445453 | 82f278907c204a7f9db510828a34a275 | K=63 | 2012.14264 | scenario_0_K_63.jpg |
0.486044 | 00b7a03c42b9470bacc71f67e5dfe306 | K=63 | 2012.14264 | scenario_0_K_63.jpg |
0.406908 | 32f0f80ab9414b39b0494f435d997075 | Beta(1, 3) prior | 2012.14264 | scenario_1.jpg |
0.415174 | d234d091c31348a687c4cbcb31850371 | Beta(1, 3) prior | 2012.14264 | scenario_1.jpg |
0.39935 | ae10c9321fff426f8e292fb93d817777 | Beta(1, 3) prior | 2012.14264 | scenario_1.jpg |
0.439054 | 9d7a78e5c708441cbc2993382079f6b0 | Scenario 1 / K=250 | 2012.14264 | scenario_1_K_250.jpg |
0.414507 | 0886d1cfeab44a159ee02566f2e8a43f | Scenario 1 / K=250 | 2012.14264 | scenario_1_K_250.jpg |
0.397208 | b7921fbe1a8847a1899188f2483bfaef | Scenario 1 / K=5 | 2012.14264 | scenario_1_K_5.jpg |
0.43966 | 3e7ad8399a6d489390da2fea6d0927e2 | Scenario 1 / K=5 | 2012.14264 | scenario_1_K_5.jpg |
0.477088 | 1aed047e497245c3887767130ab1d5b0 | Scenario 1 / K=63 | 2012.14264 | scenario_1_K_63.jpg |
0.477493 | a605b7c6ccb441889bc360e140b53806 | Scenario 1 / K=63 | 2012.14264 | scenario_1_K_63.jpg |
0.392786 | 34df91a267104385820b43968d5df612 | Bernoulli bandits with Beta(1, 3) prior | 2012.14264 | scenario_2.jpg |
0.472764 | 9f0649c3b37a4659a5657daafcff4284 | Scenario 2 / K=250 | 2012.14264 | scenario_2_K_250.jpg |
0.507457 | fd4dbea4290a4910be11aff84001cb50 | Scenario 2 / K=250 | 2012.14264 | scenario_2_K_250.jpg |
0.389972 | b5a5cf3eb27843378cc52fa0230f7882 | Scenario 2 / K=5 | 2012.14264 | scenario_2_K_5.jpg |
0.476222 | 15b153ef01874ba8bc61c21d7d4a91c7 | Scenario 2 / K=5 | 2012.14264 | scenario_2_K_5.jpg |
0.435136 | e91b2984111742ccac68c58fdca67904 | Scenario 2 / K=63 | 2012.14264 | scenario_2_K_63.jpg |
0.508469 | 5ee06758b74d457caa3721462bccc7a5 | Scenario 2 / K=63 | 2012.14264 | scenario_2_K_63.jpg |
0.500432 | 1649ff5d569043d993d0eacace04c11f | Intensity calibration procedure. (a) The average intensity profile as a function of distance, I(x) taken with laser profilometry after various n_s. Inset: The average intensity, ⟨ I ⟩, from x=0.5 to x=0.6 m, as a function of the number of sprays n_s, is well described by a linear fit with slope 2.56. (b) Matching the field image profilometry with laser profilometry. Field image profile is multiplied by a factor of 1.6 so that the maximum intensities are nearly equal. (c) The integrated intensity over the deposited area shown in Fig. <ref>(b). Integrated intensity and the volume of droplets have a linear relationship with slope 6.21 × 10^6. Volume per spray is V_s=0.75 mL such that V_20 = n_s V_s ≈ 15 mL. | 2012.14268 | App1_intensityCalibration.jpg |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.