MiDaS v3.1 - Githubissues

isl-org / MiDaS

Code for robust monocular depth estimation described in "Ranftl et. al., Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer, TPAMI 2022"

MIT License

4.27k stars 599 forks source link

MiDaS v3.1 #191

Closed rbirkl closed 1 year ago

rbirkl commented 1 year ago

New models based on 5 different types of transformers (BEiT, Swin2, Swin, Next-ViT, LeViT)
Training datasets extended from 10 to 12, including also KITTI and NYU Depth V2 using BTS split
Best model, BEiT L-512, with resolution 512x512, is on average about 28% more accurate than MiDaS 3.0
Integrated live depth estimation from camera feed