Project: SpaceNet_Off_Nadir_Solutions (GitHub Link)

SpaceNet_Off_Nadir_Solutions-master
- cannab
  - split_folds.py
  - train92_9ch_fold.py
  - train.sh
  - predict154_9ch_oof.py
  - merge_oof.py
  - predict92_9ch_oof.py
  - create_submission_lgbm.py
  - train154_9ch_fold.py
  - predict50_9ch_oof.py
  - predict92_9ch_fold.py
  - test.sh
  - zoo
    - dpn.py
    - models.py
    - __init__.py
    - senet.py
  - predict50_9ch_fold.py
  - convert_test.py
  - train50_9ch_fold.py
  - train101_9ch_fold.py
  - losses.py
  - predict_classifier.py
  - download_models.sh
  - predict154_9ch_fold.py
  - create_masks.py
  - README.md
  - utils.py
  - Dockerfile
  - train_classifier.py
  - Instructions.txt
  - adamw.py
  - merge.py
  - predict101_9ch_fold.py
  - predict101_9ch_oof.py
- selim_sef
  - generate_polygons.py
  - predict_oof.sh
  - training
    - metric.py
    - meters.py
    - eval.py
    - __pycache__
      - __init__.cpython-36.pyc
      - meters.cpython-36.pyc
      - metric.cpython-36.pyc
      - utils.cpython-36.pyc
      - losses.cpython-36.pyc
      - eval.cpython-36.pyc
    - losses.py
    - __init__.py
    - utils.py
  - train.sh
  - __pycache__
    - train_classifier.cpython-36.pyc
  - weights
    - .gitkeep
  - ensemble.py
  - train_all.sh
  - dataset
    - dense_data.py
    - __pycache__
      - __init__.cpython-36.pyc
      - dense_data.cpython-36.pyc
      - dense_transform.cpython-36.pyc
    - dense_transform.py
    - __init__.py
  - test.sh
  - zoo
    - dpn.py
    - __pycache__
      - dpn.cpython-36.pyc
      - resnet.cpython-36.pyc
      - __init__.cpython-36.pyc
      - senet.cpython-36.pyc
      - densenet.cpython-36.pyc
      - unet.cpython-36.pyc
    - resnet.py
    - densenet.py
    - unet.py
    - __init__.py
    - senet.py
  - create_folds.py
  - predict_trees.py
  - train.py
  - configs
    - r101.json
    - d121.json
    - d161.json
    - r34.json
    - sc50.json
  - README.md
  - predict.sh
  - lgbm_models
    - gbm_model_0_22.txt
    - gbm_model_0_19.txt
    - gbm_model_0_6.txt
    - gbm_model_0_17.txt
    - gbm_model_0_23.txt
    - gbm_model_0_27.txt
    - gbm_model_0_20.txt
    - gbm_model_0_26.txt
    - gbm_model_0_2.txt
  - train_labels
    - .gitkeep
  - tools
    - clr.py
    - __pycache__
      - __init__.cpython-36.pyc
      - lr_policy.cpython-36.pyc
      - adamw.cpython-36.pyc
      - config.cpython-36.pyc
      - clr.cpython-36.pyc
    - config.py
    - mask_from_geo.py
    - mask_utils.py
    - lr_policy.py
    - __init__.py
    - rle.py
    - adamw.py
  - logs
    - .gitkeep
  - inference
    - tta.py
    - predict_oof.py
    - __pycache__
      - tta.cpython-36.pyc
      - __init__.cpython-36.pyc
      - predict.cpython-36.pyc
    - predict.py
    - __init__.py
  - .idea
    - modules.xml
    - encodings.xml
    - misc.xml
    - workspace.xml
    - vcs.xml
    - spacenet.iml
  - Dockerfile
  - train_classifier.py
  - evaluate_labels.py
- MaksimovKA_solution
  - utils
    - __pycache__
      - losses.cpython-35.pyc
      - metrics.cpython-35.pyc
      - metrics.cpython-36.pyc
      - losses.cpython-36.pyc
    - metrics.py
    - losses.py
  - default_folds_split
  - train.sh
  - dataset
    - spacenet_binary_dataset.py
    - base_dataset.py
  - test.sh
  - prepare_data
    - convert_data.py
    - create_folds.py
  - models
    - qubvel_segmentation_models
      - pspnet
        __pycache__
        builder.cpython-35.pyc
        model.cpython-35.pyc
        __init__.cpython-35.pyc
        blocks.cpython-35.pyc
        blocks.py
        model.py
        __init__.py
        builder.py
      - __pycache__
        __init__.cpython-36.pyc
        utils.cpython-36.pyc
        __init__.cpython-35.pyc
        utils.cpython-35.pyc
      - common
        __pycache__
        layers.cpython-35.pyc
        __init__.cpython-35.pyc
        functions.cpython-35.pyc
        blocks.cpython-35.pyc
        blocks.py
        functions.py
        layers.py
        __init__.py
      - linknet
        __pycache__
        builder.cpython-35.pyc
        model.cpython-35.pyc
        __init__.cpython-35.pyc
        blocks.cpython-35.pyc
        blocks.py
        model.py
        __init__.py
        builder.py
      - fpn
        __pycache__
        blocks.cpython-36.pyc
        layers.cpython-35.pyc
        builder.cpython-35.pyc
        builder.cpython-36.pyc
        __init__.cpython-36.pyc
        layers.cpython-36.pyc
        model.cpython-35.pyc
        __init__.cpython-35.pyc
        blocks.cpython-35.pyc
        model.cpython-36.pyc
        blocks.py
        model.py
        layers.py
        __init__.py
        builder.py
      - __init__.py
      - backbones
        inception_v3.py
        __pycache__
        inception_v3.cpython-36.pyc
        backbones.cpython-36.pyc
        __init__.cpython-36.pyc
        inception_resnet_v2.cpython-36.pyc
        __init__.cpython-35.pyc
        inception_resnet_v2.cpython-35.pyc
        inception_v3.cpython-35.pyc
        backbones.cpython-35.pyc
        backbones.py
        inception_resnet_v2.py
        __init__.py
        classification_models
        imgs
        graphs
        tests
        __init__.py
        README.md
        tests
        __init__.py
        test_imagenet.py
        classification_models
        resnet
        blocks.py
        preprocessing.py
        models.py
        params.py
        __init__.py
        builder.py
        __init__.py
        utils.py
        weights.py
        resnext
        blocks.py
        preprocessing.py
        models.py
        params.py
        __init__.py
        builder.py
        .gitignore
      - utils.py
      - unet
        __pycache__
        blocks.cpython-36.pyc
        builder.cpython-35.pyc
        builder.cpython-36.pyc
        __init__.cpython-36.pyc
        model.cpython-35.pyc
        __init__.cpython-35.pyc
        blocks.cpython-35.pyc
        model.cpython-36.pyc
        blocks.py
        model.py
        __init__.py
        builder.py
  - predict
    - predict_segmentation.py
  - params
    - params.py
  - README.md
  - augmentations
    - transforms.py
  - train
    - train_segmentation.py
    - model_factory.py
  - Dockerfile
  - .dockerignore
- LICENSE
- number13
  - number13
    - src
      - Mask_RCNN
        LICENSE
        setup.py
        setup.cfg
        README.md
        mrcnn
        __pycache__
        model_mod_ohem.cpython-35.pyc
        model_mod.cpython-36.pyc
        config.cpython-35.pyc
        visualize.cpython-35.pyc
        model_det.cpython-35.pyc
        model_mod_mpan.cpython-36.pyc
        resnets.cpython-35.pyc
        model_resnet152.cpython-35.pyc
        visualize.cpython-36.pyc
        __init__.cpython-36.pyc
        model_mod_rgb.cpython-36.pyc
        model.cpython-35.pyc
        parallel_model.cpython-36.pyc
        model_mod.cpython-35.pyc
        utils.cpython-36.pyc
        __init__.cpython-35.pyc
        config.cpython-36.pyc
        model_mod_alldil.cpython-35.pyc
        utils.cpython-35.pyc
        config.py
        parallel_model.py
        __init__.pyc
        model.py
        __init__.py
        utils.py
        visualize.py
        utils.pyc
        model_mod_rgb.py
        model_mod_mpan.py
        requirements.txt
        logs
        MANIFEST.in
      - util.py
      - eval_val.py
      - patchify.py
      - train.sh
      - config.py
      - prediction.py
      - test.sh
      - create_patches_all.py
      - models.py
      - crowdai_train.sh
      - crowdai_train.py
      - cocoeval.py
      - train.py
      - inference.py
      - val_set.pkl
    - README.txt
    - Dockerfile
  - README.md
- README.md
- XD_XD
  - train.sh
  - working
  - INSTRUCTION.md
  - test.sh
  - main.py
  - README.md
  - Dockerfile
  - .dockerignore
- .gitignore

SpaceNet Challenge Round 4: Off-Nadir Buildings

Competitor solutions

Summary

The five subdirectories in this repository comprise the code for the winning solutions of SpaceNet 4: Off-Nadir Building Detection hosted by TopCoder. Each subdirectory contains the competitors' written descriptions of their solution to the challenge. See the blog post on CosmiQ Works' blog The DownlinQ for an additional summary.

Coming soon

Links to download the competitors' trained models

Questions about SpaceNet? Check out our website at https://spacenet.ai.