Single Image Depth Estimation Trained via Depth from Defocus Cues

Benchmark Model Rank Results
monocular-depth-estimation-on-kitti-eigen-splitDeepLabV3+ (F10)#56absolute relative error: 0.110
monocular-depth-estimation-on-nyu-depth-v2DeepLabV3+ (F10)#75RMSE: 0.575