Tutorial 3: Custom Data Pipelines

Design of Data pipelines

Following typical conventions, we use Dataset and DataLoader for data loading with multiple workers. Dataset returns a dict of data items corresponding the arguments of models’ forward method. Since the data in object detection may not be the same size (image size, gt bbox size, etc.), we introduce a new DataContainer type in MMCV to help collect and distribute data of different size. See here for more details.

The data preparation pipeline and the dataset is decomposed. Usually a dataset defines how to process the annotations and a data pipeline defines all the steps to prepare a data dict. A pipeline consists of a sequence of operations. Each operation takes a dict as input and also output a dict for the next transform.

The operations are categorized into data loading, pre-processing, formatting, label generating.

Here is an pipeline example for Simple Baseline (ResNet50).

train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='TopDownRandomFlip', flip_prob=0.5),
    dict(type='TopDownHalfBodyTransform', num_joints_half_body=8, prob_half_body=0.3),
    dict(type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
    dict(type='TopDownAffine'),
    dict(type='ToTensor'),
    dict(
        type='NormalizeTensor',
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225]),
    dict(type='TopDownGenerateTarget', sigma=2),
    dict(
        type='Collect',
        keys=['img', 'target', 'target_weight'],
        meta_keys=[
            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
            'rotation', 'bbox_score', 'flip_pairs'
        ]),
]

val_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='TopDownAffine'),
    dict(type='ToTensor'),
    dict(
        type='NormalizeTensor',
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225]),
    dict(
        type='Collect',
        keys=['img'],
        meta_keys=[
            'image_file', 'center', 'scale', 'rotation', 'bbox_score',
            'flip_pairs'
        ]),
]

For each operation, we list the related dict fields that are added/updated/removed.

Data loading

LoadImageFromFile

  • add: img, img_file

Pre-processing

TopDownRandomFlip

  • update: img, joints_3d, joints_3d_visible, center

TopDownHalfBodyTransform

  • update: center, scale

TopDownGetRandomScaleRotation

  • update: scale, rotation

TopDownAffine

  • update: img, joints_3d, joints_3d_visible

NormalizeTensor

  • update: img

Generating labels

TopDownGenerateTarget

  • add: target, target_weight

Formatting

ToTensor

  • update: ‘img’

Collect

  • add: img_meta (the keys of img_meta is specified by meta_keys)
  • remove: all other keys except for those specified by keys

Extend and use custom pipelines

  1. Write a new pipeline in any file, e.g., my_pipeline.py. It takes a dict as input and return a dict.

    from mmpose.datasets import PIPELINES
    
    @PIPELINES.register_module()
    class MyTransform:
    
       def __call__(self, results):
           results['dummy'] = True
           return results
    
  2. Import the new class.

    from .my_pipeline import MyTransform
    
  3. Use it in config files.

    train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='TopDownRandomFlip', flip_prob=0.5),
    dict(type='TopDownHalfBodyTransform', num_joints_half_body=8, prob_half_body=0.3),
    dict(type='TopDownGetRandomScaleRotation', rot_factor=40, scale_factor=0.5),
    dict(type='TopDownAffine'),
    dict(type='MyTransform'),
    dict(type='ToTensor'),
    dict(
        type='NormalizeTensor',
        mean=[0.485, 0.456, 0.406],
        std=[0.229, 0.224, 0.225]),
    dict(type='TopDownGenerateTarget', sigma=2),
    dict(
        type='Collect',
        keys=['img', 'target', 'target_weight'],
        meta_keys=[
            'image_file', 'joints_3d', 'joints_3d_visible', 'center', 'scale',
            'rotation', 'bbox_score', 'flip_pairs'
        ]),
    ]