Monocular Depth Estimation Survey

In this repo, we provide a unified framework for Monocular Depth Estimation.

Our framework provides a convenient interface for various dataset and various methods, which supports a fair comparison by aligning the output and evaluation scripts.

Evaluation

The dataset, model and evaluation metric configuration can be set in the yaml file in configs. E.g.,

• Dataset Config:

  dataset: NYUv2
  dataset_params:
    path: /mnt/pfs/data/RGBD/moge_eval/NYUv2
    width: 640
    height: 480
    split: ".index.txt"
    depth_unit: 1.0
  

• Model Config:

  model_name: "Marigold"
  model_params:
    model_dir: "/mnt/pfs/users/sunyangtian/Depth/Marigold"
    ckpt_path: "/mnt/pfs/share/pretrained_model/marigold-depth-v1-1"
    denoise_steps: 1
    ensemble_size: 1
    half_precision: False
    processing_res: 0
    output_processing_res: False
    resample_method: bilinear
    color_map: Spectral
  

• Metric Config

  eval_depth:
    metric_names: 
      - 'Abs Rel'
      - 'delta < 1.25'
      - 'delta < 1.25^2'
      - 'delta < 1.25^3'
    depth_alignment: "lstsq"
    metric_scale: False
  

• Output Config

  vis_depth: True
  save_dir: debug_marigold_nyuv2
  

Finally, the evaluation process can be performed by

  python eval.py configs/moge_benchmark/marigold/marigold_nyuv2.yaml

You can also use eval_all.sh to evaluate all datasets with one command.

Supported Datasets

Please refer to dataset for more details.

Supported Methods

Please refer to model for more details.

Acknowledgements