Delta < 1.25: 0.4324Delta < 1.25^2: 0.6667Delta < 1.25^3: 0.7949Average PSNR: 29.2636
Papers with Code · CC BY-SA 4.0
Depth Estimation
Depth Estimation is the task of measuring the distance of each pixel relative to the camera. Depth is extracted from either monocular (single) or stereo (multiple views of a scene) images. Traditional methods use multi-view geometry to find the relationship between the images. Newer methods can directly estimate depth by minimizing the regression loss, or by learning to generate a novel view from a sequence. The most popular benchmarks are KITTI and NYUv2. Models are typically evaluated according to a RMS metric.
<span class="description-source">Source: DIODE: A Dense Indoor and Outdoor DEpth Dataset </span>
논문
RMSE: 0.165
RMS: 0.4325
Absolute relative error (AbsRel): 0.076Sq Rel: 0.607RMSE: 0.04025
RMS: 0.287
RMS: 0.30
A2J: Anchor-to-Joint Regression Network for 3D Articulated Pose Estimation from a Single Depth Image
mAP: 8.61
RMS: 0.364
Delta < 1.25^3: 0.7945Delta < 1.25: 0.3563Delta < 1.25^2: 0.5948
BadPix(0.07): 3.756MSE : 1.904BadPix(0.03): 6.823BadPix(0.01): 17.226
RMS: 0.509
RMS: 0.279
absolute relative error: 0.043Delta < 1.25: 0.981
RMS: 0.416
RMS: 0.224
RMSE: 0.013
RMS: 0.407
RMSE: 0.128absolute relative error: 0.026Delta < 1.25: 0.988Delta < 1.25^2: 1.000
Delta < 1.25: 0.234
Number of parameters (M): 42.6
RMS: 0.586
RMS: 0.356
RMS: 0.792
L1 error: 1.63
RMS: 0.365