Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Support keypoint on ElasticTransform (2-D) #1325

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
Open

Support keypoint on ElasticTransform (2-D) #1325

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
71aefd4
Support keypoint on ``ElasticTransform`` (2-D)
minipuding Oct 23, 2022
7d72b84
Modified.. keypoints of ``ElasticTransform``.
minipuding Oct 26, 2022
c2a92d9
Add files via upload
minipuding Oct 26, 2022
ede9e5e
Modified.. keypoints of ElasticTransform.
minipuding Oct 26, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
27 changes: 25 additions & 2 deletions albumentations/augmentations/geometric/transforms.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
import math
import random
from enum import Enum
from typing import Dict, Optional, Sequence, Tuple, Union
from typing import Dict, Optional, Sequence, Tuple, Union, List

import cv2
import numpy as np
Expand Down Expand Up @@ -203,12 +203,13 @@ def __init__(
self.same_dxdy = same_dxdy

def apply(self, img, random_state=None, interpolation=cv2.INTER_LINEAR, **params):
# For supporting ``apply_to_keypoint`` function, Using ``elastic_transform_v2`` to replace ``elastic_transform``
return F.elastic_transform(
img,
self.alpha,
self.sigma,
self.alpha_affine,
interpolation,
self.interpolation,
self.border_mode,
self.value,
np.random.RandomState(random_state),
Expand Down Expand Up @@ -251,6 +252,28 @@ def apply_to_bbox(self, bbox, random_state=None, **params):
bbox_returned = bbox_from_mask(mask)
bbox_returned = F.normalize_bbox(bbox_returned, rows, cols)
return bbox_returned

def apply_to_keypoint(self, keypoint: KeypointInternalType, random_state=None, half_win=9, **params) -> KeypointInternalType:
"""
Only consider 2-D case.
"""
image = np.zeros([params["cols"], params["rows"]])
X, Y = np.meshgrid(np.arange(image.shape[0]), np.arange(image.shape[1]))
image[Y, X] = np.exp(- 0.5 / (0.02 ** 2) *
(((X - keypoint[0]) / image.shape[0]) ** 2 + ((Y - keypoint[1]) / image.shape[1]) ** 2))
remap_image = F.elastic_transform(
image,
self.alpha,
self.sigma,
self.alpha_affine,
self.interpolation,
self.border_mode,
self.mask_value,
np.random.RandomState(random_state),
self.approximate,
)
interp_y, interp_x = np.where(remap_image == np.max(remap_image))
return (interp_x[0], interp_y[0], 0.0, 0.0)
Comment on lines +275 to +276
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What if the keypoint outside of the image? Looks like we always will find keypoint inside the image and in the case when we can not see the keypoint

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, there is such a problem, but I can't think of a simple and elegant way to solve it. Therefore, you should try to ensure that the keypoints do not go beyond the scope of the image, otherwise it will return incorrect results.


def get_params(self):
return {"random_state": random.randint(0, 10000)}
Expand Down
43 changes: 6 additions & 37 deletions albumentations/augmentations/transforms.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,7 +222,7 @@ def get_transform_init_args_names(self):


class ImageCompression(ImageOnlyTransform):
"""Decreases image quality by Jpeg, WebP compression of an image.
"""Decrease Jpeg, WebP compression of an image.

Args:
quality_lower (float): lower bound on the image quality.
Expand Down Expand Up @@ -292,7 +292,7 @@ def get_transform_init_args(self):


class JpegCompression(ImageCompression):
"""Decreases image quality by Jpeg compression of an image.
"""Decrease Jpeg compression of an image.

Args:
quality_lower (float): lower bound on the jpeg quality. Should be in [0, 100] range
Expand Down Expand Up @@ -2391,10 +2391,6 @@ class Spatter(ImageOnlyTransform):
If tuple of float intensity will be sampled from range `[intensity[0], intensity[1])`. Default: (0.6).
mode (string, or list of strings): Type of corruption. Currently, supported options are 'rain' and 'mud'.
If list is provided type of corruption will be sampled list. Default: ("rain").
color (list of (r, g, b) or dict or None): Corruption elements color.
If list uses provided list as color for specified mode.
If dict uses provided color for specified mode. Color for each specified mode should be provided in dict.
If None uses default colors (rain: (238, 238, 175), mud: (20, 42, 63)).
p (float): probability of applying the transform. Default: 0.5.

Targets:
Expand All @@ -2416,7 +2412,6 @@ def __init__(
cutout_threshold: ScaleFloatType = 0.68,
intensity: ScaleFloatType = 0.6,
mode: Union[str, Sequence[str]] = "rain",
color: Optional[Union[Sequence[int], Dict[str, Sequence[int]]]] = None,
always_apply: bool = False,
p: float = 0.5,
):
Expand All @@ -2427,34 +2422,10 @@ def __init__(
self.gauss_sigma = to_tuple(gauss_sigma, gauss_sigma)
self.intensity = to_tuple(intensity, intensity)
self.cutout_threshold = to_tuple(cutout_threshold, cutout_threshold)
self.color = (
color
if color is not None
else {
"rain": [238, 238, 175],
"mud": [20, 42, 63],
}
)
self.mode = mode if isinstance(mode, (list, tuple)) else [mode]

if len(set(self.mode)) > 1 and not isinstance(self.color, dict):
raise ValueError(f"Unsupported color: {self.color}. Please specify color for each mode (use dict for it).")

for i in self.mode:
if i not in ["rain", "mud"]:
raise ValueError(f"Unsupported color mode: {mode}. Transform supports only `rain` and `mud` mods.")
if isinstance(self.color, dict):
if i not in self.color:
raise ValueError(f"Wrong color definition: {self.color}. Color for mode: {i} not specified.")
if len(self.color[i]) != 3:
raise ValueError(
f"Unsupported color: {self.color[i]} for mode {i}. Color should be presented in RGB format."
)

if isinstance(self.color, (list, tuple)):
if len(self.color) != 3:
raise ValueError(f"Unsupported color: {self.color}. Color should be presented in RGB format.")
self.color = {self.mode[0]: self.color}

def apply(
self,
Expand All @@ -2480,7 +2451,6 @@ def get_params_dependent_on_targets(self, params: Dict[str, Any]) -> Dict[str, A
sigma = random.uniform(self.gauss_sigma[0], self.gauss_sigma[1])
mode = random.choice(self.mode)
intensity = random.uniform(self.intensity[0], self.intensity[1])
color = np.array(self.color[mode]) / 255.0

liquid_layer = random_utils.normal(size=(h, w), loc=mean, scale=std)
liquid_layer = gaussian_filter(liquid_layer, sigma=sigma, mode="nearest")
Expand All @@ -2501,16 +2471,15 @@ def get_params_dependent_on_targets(self, params: Dict[str, Any]) -> Dict[str, A
m = liquid_layer * dist
m *= 1 / np.max(m, axis=(0, 1))

drops = m[:, :, None] * color * intensity
drops = m[:, :, None] * np.array([238 / 255.0, 238 / 255.0, 175 / 255.0]) * intensity
mud = None
non_mud = None
else:
m = np.where(liquid_layer > cutout_threshold, 1, 0)
m = gaussian_filter(m.astype(np.float32), sigma=sigma, mode="nearest")
m[m < 1.2 * cutout_threshold] = 0
m = m[..., np.newaxis]

mud = m * color
mud = m * np.array([20 / 255.0, 42 / 255.0, 63 / 255.0])
non_mud = 1 - m
drops = None

Expand All @@ -2521,5 +2490,5 @@ def get_params_dependent_on_targets(self, params: Dict[str, Any]) -> Dict[str, A
"mode": mode,
}

def get_transform_init_args_names(self) -> Tuple[str, str, str, str, str, str, str]:
return "mean", "std", "gauss_sigma", "intensity", "cutout_threshold", "mode", "color"
def get_transform_init_args_names(self) -> Tuple[str, str, str, str, str, str]:
return "mean", "std", "gauss_sigma", "intensity", "cutout_threshold", "mode"