SCIPaD: Incorporating Spatial Clues into Unsupervised Pose-Depth Joint Learning

Benchmark Model Rank Results
camera-pose-estimation-on-kitti-odometrySCIPaD#2Average Translational Error et[%]: 8.63
monocular-depth-estimation-on-kitti-eigen-1SCIPaD#4absolute relative error: 0.090RMSE: 4.056RMSE log: 0.166
monocular-depth-estimation-on-kitti-eigen-1SCIPaD(M+640x192)#19absolute relative error: 0.098RMSE: 4.391RMSE log: 0.175