Skip to content

Latest commit

 

History

History
199 lines (155 loc) · 5.4 KB

add_datasets.md

File metadata and controls

199 lines (155 loc) · 5.4 KB
Raw

Add New Datasets

Add new custom dataset

Here we show how to develop a new custom dataset.

  1. Create a new file mmseg/datasets/example.py

    from mmseg.registry import DATASETS
from .basesegdataset import BaseSegDataset


@DATASETS.register_module()
class ExampleDataset(BaseSegDataset):

    METAINFO = dict(
        classes=('xxx', 'xxx', ...),
        palette=[[x, x, x], [x, x, x], ...])

    def __init__(self, arg1, arg2):
        pass

  2. Import the module in mmseg/datasets/__init__.py

    from .example import ExampleDataset

  3. Use it by creating a new new dataset config file configs/_base_/datasets/example_dataset.py

    dataset_type = 'ExampleDataset'
data_root = 'data/example/'
...

  4. Add dataset meta information in mmseg/utils/class_names.py

    def example_classes():
    return [
        'xxx', 'xxx',
        ...
    ]

def example_palette():
    return [
        [x, x, x], [x, x, x],
        ...
    ]
dataset_aliases ={
    'example': ['example', ...],
    ...
}

Note: If the new dataset does not satisfy the mmseg requirements, a data preprocessing script needs to be prepared in tools/dataset_converters/

Customize datasets by reorganizing data

The simplest way is to convert your dataset to organize your data into folders.

An example of file structure is as followed.

├── data
│   ├── my_dataset
│   │   ├── img_dir
│   │   │   ├── train
│   │   │   │   ├── xxx{img_suffix}
│   │   │   │   ├── yyy{img_suffix}
│   │   │   │   ├── zzz{img_suffix}
│   │   │   ├── val
│   │   ├── ann_dir
│   │   │   ├── train
│   │   │   │   ├── xxx{seg_map_suffix}
│   │   │   │   ├── yyy{seg_map_suffix}
│   │   │   │   ├── zzz{seg_map_suffix}
│   │   │   ├── val

A training pair will consist of the files with same suffix in img_dir/ann_dir.

Some datasets don't release the test set or don't release the ground truth of the test set, and we cannot evaluate models locally without the ground truth of the test set, so we set the validation set as the default test set in config files.

About how to build your own datasets or implement a new dataset class please refer to the datasets guide for more detailed information.

Note: The annotations are images of shape (H, W), the value pixel should fall in range [0, num_classes - 1]. You may use 'P' mode of pillow to create your annotation image with color.

Customize datasets by mixing dataset

MMSegmentation also supports to mix dataset for training. Currently it supports to concat, repeat and multi-image mix datasets.

Repeat dataset

We use RepeatDataset as wrapper to repeat the dataset. For example, suppose the original dataset is Dataset_A, to repeat it, the config looks like the following

dataset_A_train = dict(
    type='RepeatDataset',
    times=N,
    dataset=dict(  # This is the original config of Dataset_A
        type='Dataset_A',
        ...
        pipeline=train_pipeline
    )
)

Concatenate dataset

In case the dataset you want to concatenate is different, you can concatenate the dataset configs like the following.

dataset_A_train = dict()
dataset_B_train = dict()
concatenate_dataset = dict(
    type='ConcatDataset',
    datasets=[dataset_A_train, dataset_B_train])

A more complex example that repeats Dataset_A and Dataset_B by N and M times, respectively, and then concatenates the repeated datasets is as the following.

dataset_A_train = dict(
    type='RepeatDataset',
    times=N,
    dataset=dict(
        type='Dataset_A',
        ...
        pipeline=train_pipeline
    )
)
dataset_A_val = dict(
    ...
    pipeline=test_pipeline
)
dataset_A_test = dict(
    ...
    pipeline=test_pipeline
)
dataset_B_train = dict(
    type='RepeatDataset',
    times=M,
    dataset=dict(
        type='Dataset_B',
        ...
        pipeline=train_pipeline
    )
)
train_dataloader = dict(
    dataset=dict(
        type='ConcatDataset',
        datasets=[dataset_A_train, dataset_B_train]))

val_dataloader = dict(dataset=dataset_A_val)
test_dataloader = dict(dataset=dataset_A_test)

You can refer base dataset tutorial from mmengine for more details

Multi-image Mix Dataset

We use MultiImageMixDataset as a wrapper to mix images from multiple datasets. MultiImageMixDataset can be used by multiple images mixed data augmentation like mosaic and mixup.

An example of using MultiImageMixDataset with Mosaic data augmentation:

train_pipeline = [
    dict(type='RandomMosaic', prob=1),
    dict(type='Resize', img_scale=(1024, 512), keep_ratio=True),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PackSegInputs')
]

train_dataset = dict(
    type='MultiImageMixDataset',
    dataset=dict(
        type=dataset_type,
        reduce_zero_label=False,
        img_dir=data_root + "images/train",
        ann_dir=data_root + "annotations/train",
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(type='LoadAnnotations'),
        ]
    ),
    pipeline=train_pipeline
)