mmsegmentation/tests/test_datasets/test_loading.py

279 lines
10 KiB
Python
Raw Normal View History

# Copyright (c) OpenMMLab. All rights reserved.
import copy
import os.path as osp
import tempfile
import mmcv
import numpy as np
from mmcv.transforms import LoadImageFromFile
from mmseg.datasets.transforms import (LoadAnnotations,
LoadBiomedicalAnnotation,
LoadBiomedicalData,
LoadBiomedicalImageFromFile,
LoadImageFromNDArray)
class TestLoading:
@classmethod
def setup_class(cls):
cls.data_prefix = osp.join(osp.dirname(__file__), '../data')
def test_load_img(self):
results = dict(img_path=osp.join(self.data_prefix, 'color.jpg'))
transform = LoadImageFromFile()
results = transform(copy.deepcopy(results))
assert results['img_path'] == osp.join(self.data_prefix, 'color.jpg')
assert results['img'].shape == (288, 512, 3)
assert results['img'].dtype == np.uint8
assert results['ori_shape'] == results['img'].shape[:2]
assert repr(transform) == transform.__class__.__name__ + \
2022-06-27 22:36:18 +08:00
"(ignore_empty=False, to_float32=False, color_type='color'," + \
" imdecode_backend='cv2', backend_args=None)"
# to_float32
transform = LoadImageFromFile(to_float32=True)
results = transform(copy.deepcopy(results))
assert results['img'].dtype == np.float32
# gray image
results = dict(img_path=osp.join(self.data_prefix, 'gray.jpg'))
transform = LoadImageFromFile()
results = transform(copy.deepcopy(results))
assert results['img'].shape == (288, 512, 3)
assert results['img'].dtype == np.uint8
transform = LoadImageFromFile(color_type='unchanged')
results = transform(copy.deepcopy(results))
assert results['img'].shape == (288, 512)
assert results['img'].dtype == np.uint8
def test_load_seg(self):
seg_path = osp.join(self.data_prefix, 'seg.png')
results = dict(
seg_map_path=seg_path, reduce_zero_label=True, seg_fields=[])
transform = LoadAnnotations()
results = transform(copy.deepcopy(results))
assert results['gt_seg_map'].shape == (288, 512)
assert results['gt_seg_map'].dtype == np.uint8
assert repr(transform) == transform.__class__.__name__ + \
"(reduce_zero_label=True, imdecode_backend='pillow', " + \
'backend_args=None)'
# reduce_zero_label
transform = LoadAnnotations(reduce_zero_label=True)
results = transform(copy.deepcopy(results))
assert results['gt_seg_map'].shape == (288, 512)
assert results['gt_seg_map'].dtype == np.uint8
def test_load_seg_custom_classes(self):
test_img = np.random.rand(10, 10)
test_gt = np.zeros_like(test_img)
test_gt[2:4, 2:4] = 1
test_gt[2:4, 6:8] = 2
test_gt[6:8, 2:4] = 3
test_gt[6:8, 6:8] = 4
tmp_dir = tempfile.TemporaryDirectory()
img_path = osp.join(tmp_dir.name, 'img.jpg')
gt_path = osp.join(tmp_dir.name, 'gt.png')
mmcv.imwrite(test_img, img_path)
mmcv.imwrite(test_gt, gt_path)
# test only train with label with id 3
results = dict(
img_path=img_path,
seg_map_path=gt_path,
label_map={
0: 0,
1: 0,
2: 0,
3: 1,
4: 0
},
reduce_zero_label=False,
seg_fields=[])
load_imgs = LoadImageFromFile()
results = load_imgs(copy.deepcopy(results))
load_anns = LoadAnnotations()
results = load_anns(copy.deepcopy(results))
gt_array = results['gt_seg_map']
true_mask = np.zeros_like(gt_array)
true_mask[6:8, 2:4] = 1
assert results['seg_fields'] == ['gt_seg_map']
assert gt_array.shape == (10, 10)
assert gt_array.dtype == np.uint8
np.testing.assert_array_equal(gt_array, true_mask)
# test only train with label with id 4 and 3
results = dict(
img_path=osp.join(self.data_prefix, 'color.jpg'),
seg_map_path=gt_path,
label_map={
0: 0,
1: 0,
2: 0,
3: 2,
4: 1
},
reduce_zero_label=False,
seg_fields=[])
load_imgs = LoadImageFromFile()
results = load_imgs(copy.deepcopy(results))
load_anns = LoadAnnotations()
results = load_anns(copy.deepcopy(results))
gt_array = results['gt_seg_map']
true_mask = np.zeros_like(gt_array)
true_mask[6:8, 2:4] = 2
true_mask[6:8, 6:8] = 1
assert results['seg_fields'] == ['gt_seg_map']
assert gt_array.shape == (10, 10)
assert gt_array.dtype == np.uint8
np.testing.assert_array_equal(gt_array, true_mask)
[Fix] Switch order of `reduce_zero_label` and applying `label_map` in 1.x (#2517) This is an almost exact duplicate of #2500 (that was made to the `master` branch) now applied to the `1.x` branch. --- ## Motivation I want to fix a bug through this PR. The bug occurs when two options -- `reduce_zero_label=True`, and custom classes are used. `reduce_zero_label` remaps the GT seg labels by remapping the zero-class to 255 which is ignored. Conceptually, this should occur *before* the `label_map` is applied, which maps *already reduced labels*. However, currently, the `label_map` is applied before the zero label is reduced. ## Modification The modification is simple: - I've just interchanged the order of the two operations by moving a few lines from bottom to top. - I've added a test that passes when the fix is introduced, and fails on the original `master` branch. ## BC-breaking (Optional) I do not anticipate this change braking any backward-compatibility. ## Checklist - [x] Pre-commit or other linting tools are used to fix the potential lint issues. - _I've fixed all linting/pre-commit errors._ - [x] The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness. - _I've added a unit test._ - [x] If the modification has potential influence on downstream projects, this PR should be tested with downstream projects, like MMDet or MMDet3D. - _I don't think this change affects MMDet or MMDet3D._ - [x] The documentation has been modified accordingly, like docstring or example tutorials. - _This change fixes an existing bug and doesn't require modifying any documentation/docstring._
2023-01-30 12:17:15 +08:00
# test with removing a class and reducing zero label simultaneously
results = dict(
img_path=img_path,
seg_map_path=gt_path,
# since reduce_zero_label is True, there are only 4 real classes.
# if the full set of classes is ["A", "B", "C", "D"], the
# following label map simulates the dataset option
# classes=["A", "C", "D"] which removes class "B".
label_map={
0: 0,
1: 255, # simulate removing class 1
2: 1,
3: 2
},
reduce_zero_label=True, # reduce zero label
seg_fields=[])
load_imgs = LoadImageFromFile()
results = load_imgs(copy.deepcopy(results))
# reduce zero label
load_anns = LoadAnnotations()
results = load_anns(copy.deepcopy(results))
gt_array = results['gt_seg_map']
true_mask = np.ones_like(gt_array) * 255 # all zeros get mapped to 255
true_mask[2:4, 2:4] = 0 # 1s are reduced to class 0 mapped to class 0
true_mask[2:4, 6:8] = 255 # 2s are reduced to class 1 which is removed
true_mask[6:8, 2:4] = 1 # 3s are reduced to class 2 mapped to class 1
true_mask[6:8, 6:8] = 2 # 4s are reduced to class 3 mapped to class 2
assert results['seg_fields'] == ['gt_seg_map']
assert gt_array.shape == (10, 10)
assert gt_array.dtype == np.uint8
np.testing.assert_array_equal(gt_array, true_mask)
# test no custom classes
results = dict(
img_path=img_path,
seg_map_path=gt_path,
reduce_zero_label=False,
seg_fields=[])
load_imgs = LoadImageFromFile()
results = load_imgs(copy.deepcopy(results))
load_anns = LoadAnnotations()
results = load_anns(copy.deepcopy(results))
gt_array = results['gt_seg_map']
assert results['seg_fields'] == ['gt_seg_map']
assert gt_array.shape == (10, 10)
assert gt_array.dtype == np.uint8
np.testing.assert_array_equal(gt_array, test_gt)
tmp_dir.cleanup()
def test_load_image_from_ndarray(self):
results = {'img': np.zeros((256, 256, 3), dtype=np.uint8)}
transform = LoadImageFromNDArray()
results = transform(results)
assert results['img'].shape == (256, 256, 3)
assert results['img'].dtype == np.uint8
assert results['img_shape'] == (256, 256)
assert results['ori_shape'] == (256, 256)
# to_float32
transform = LoadImageFromNDArray(to_float32=True)
results = transform(copy.deepcopy(results))
assert results['img'].dtype == np.float32
# test repr
transform = LoadImageFromNDArray()
assert repr(transform) == ('LoadImageFromNDArray('
'ignore_empty=False, '
'to_float32=False, '
"color_type='color', "
"imdecode_backend='cv2', "
'backend_args=None)')
def test_load_biomedical_img(self):
results = dict(
img_path=osp.join(self.data_prefix, 'biomedical.nii.gz'))
transform = LoadBiomedicalImageFromFile()
results = transform(copy.deepcopy(results))
assert results['img_path'] == osp.join(self.data_prefix,
'biomedical.nii.gz')
assert len(results['img'].shape) == 4
assert results['img'].dtype == np.float32
assert results['ori_shape'] == results['img'].shape[1:]
assert repr(transform) == ('LoadBiomedicalImageFromFile('
"decode_backend='nifti', "
'to_xyz=False, '
'to_float32=True, '
'backend_args=None)')
def test_load_biomedical_annotation(self):
results = dict(
seg_map_path=osp.join(self.data_prefix, 'biomedical_ann.nii.gz'))
transform = LoadBiomedicalAnnotation()
results = transform(copy.deepcopy(results))
assert len(results['gt_seg_map'].shape) == 3
assert results['gt_seg_map'].dtype == np.float32
def test_load_biomedical_data(self):
input_results = dict(
img_path=osp.join(self.data_prefix, 'biomedical.npy'))
transform = LoadBiomedicalData(with_seg=True)
results = transform(copy.deepcopy(input_results))
assert results['img_path'] == osp.join(self.data_prefix,
'biomedical.npy')
assert results['img'][0].shape == results['gt_seg_map'].shape
assert results['img'].dtype == np.float32
assert results['ori_shape'] == results['img'].shape[1:]
assert repr(transform) == ('LoadBiomedicalData('
'with_seg=True, '
"decode_backend='numpy', "
'to_xyz=False, '
'backend_args=None)')
transform = LoadBiomedicalData(with_seg=False)
results = transform(copy.deepcopy(input_results))
assert len(results['img'].shape) == 4
assert results.get('gt_seg_map') is None
assert repr(transform) == ('LoadBiomedicalData('
'with_seg=False, '
"decode_backend='numpy', "
'to_xyz=False, '
'backend_args=None)')