0
# scikit-image
1

2
A comprehensive Python library for image processing and computer vision. scikit-image provides a collection of algorithms for image processing including color space conversion, filtering, morphology, segmentation, feature detection, geometric transformations, and measurement operations. It serves as the foundational image processing library for the Python scientific ecosystem.
3

4
## Package Information
5

6
- **Package Name**: scikit-image
7
- **Language**: Python
8
- **Installation**: `pip install scikit-image`
9

10
## Core Imports
11

12
```python
13
import skimage
14
```
15

16
Common submodule imports:
17

18
```python
19
from skimage import io, filters, morphology, segmentation
20
from skimage.feature import canny, corner_harris
21
from skimage.transform import resize, rotate
22
from skimage.measure import label, regionprops
23
```
24

25
## Basic Usage
26

27
```python
28
import numpy as np
29
from skimage import io, filters, morphology, measure
30

31
# Load an image
32
image = io.imread('path/to/image.jpg')
33

34
# Convert to grayscale if needed
35
from skimage.color import rgb2gray
36
gray_image = rgb2gray(image)
37

38
# Apply Gaussian filter for noise reduction
39
smooth = filters.gaussian(gray_image, sigma=1.0)
40

41
# Edge detection
42
edges = filters.sobel(smooth)
43

44
# Morphological operations
45
cleaned = morphology.remove_small_objects(edges > 0.1, min_size=100)
46

47
# Label connected components
48
labeled = measure.label(cleaned)
49
props = measure.regionprops(labeled)
50

51
# Basic measurements
52
for prop in props:
53
    print(f"Area: {prop.area}, Centroid: {prop.centroid}")
54

55
# Save processed image
56
io.imsave('processed_image.jpg', cleaned)
57
```
58

59
## Architecture
60

61
scikit-image is organized into focused submodules, each targeting specific image processing domains:
62

63
- **Core Processing**: Filtering, morphology, and transformations for fundamental image operations
64
- **Analysis & Measurement**: Feature detection, region properties, and quantitative analysis
65
- **Color & Intensity**: Color space conversion and intensity adjustment operations
66
- **Segmentation**: Algorithms to partition images into meaningful regions
67
- **Input/Output**: Flexible image reading and writing with format support
68
- **Utilities**: Data type conversion, noise generation, and helper functions
69

70
This modular design enables selective imports and provides clear separation of concerns for different image processing workflows.
71

72
## Capabilities
73

74
### Color Space Conversion
75

76
Convert between different color representations including RGB, HSV, LAB, XYZ and specialized spaces for histological imaging and color analysis.
77

78
```python { .api }
79
def rgb2gray(rgb): ...
80
def rgb2hsv(rgb): ...  
81
def rgb2lab(rgb, illuminant='D65', observer='2'): ...
82
def hsv2rgb(hsv): ...
83
def lab2rgb(lab, illuminant='D65', observer='2'): ...
84
def rgba2rgb(rgba, background=(1, 1, 1)): ...
85
```
86

87
[Color Processing](./color.md)
88

89
### Image Filtering and Enhancement
90

91
Apply filters for noise reduction, edge detection, sharpening, and enhancement. Includes Gaussian filters, edge detection operators, ridge filters, and thresholding methods.
92

93
```python { .api }
94
def gaussian(image, sigma=1, **kwargs): ...
95
def sobel(image, mask=None, **kwargs): ...
96
def canny(image, sigma=1, low_threshold=None, **kwargs): ...
97
def threshold_otsu(image, nbins=256): ...
98
def median(image, disk=None, **kwargs): ...
99
def unsharp_mask(image, radius=1.0, amount=1.0, **kwargs): ...
100
```
101

102
[Filtering](./filtering.md)
103

104
### Morphological Operations
105

106
Perform morphological operations including erosion, dilation, opening, closing, and advanced operations with customizable structuring elements.
107

108
```python { .api }
109
def erosion(image, footprint=None, **kwargs): ...
110
def dilation(image, footprint=None, **kwargs): ...
111
def opening(image, footprint=None, **kwargs): ...
112
def closing(image, footprint=None, **kwargs): ...
113
def skeletonize(image, method='lee'): ...
114
def remove_small_objects(ar, min_size=64, **kwargs): ...
115
```
116

117
[Morphology](./morphology.md)
118

119
### Geometric Transformations
120

121
Transform images through rotation, scaling, warping, and coordinate system transformations. Includes support for various transformation models and image pyramids.
122

123
```python { .api }
124
def resize(image, output_shape, **kwargs): ...
125
def rotate(image, angle, **kwargs): ...
126
def rescale(image, scale, **kwargs): ...
127
def warp(image, inverse_map, **kwargs): ...
128
class AffineTransform(matrix=None, **kwargs): ...
129
class ProjectiveTransform(matrix=None, **kwargs): ...
130
```
131

132
[Transformations](./transform.md)
133

134
### Image Segmentation
135

136
Partition images into regions using watershed, region growing, active contours, and superpixel algorithms for object detection and analysis.
137

138
```python { .api }
139
def watershed(image, markers=None, **kwargs): ...
140
def random_walker(data, labels, **kwargs): ...
141
def slic(image, n_segments=100, **kwargs): ...
142
def felzenszwalb(image, scale=1, **kwargs): ...
143
def chan_vese(image, **kwargs): ...
144
def active_contour(image, snake, **kwargs): ...
145
```
146

147
[Segmentation](./segmentation.md)
148

149
### Feature Detection and Description
150

151
Detect and describe image features including corners, edges, blobs, and keypoints. Includes modern feature descriptors like SIFT, ORB, and texture analysis methods.
152

153
```python { .api }
154
def corner_harris(image, method='k', **kwargs): ...
155
def corner_peaks(image, **kwargs): ...
156
def blob_log(image, **kwargs): ...
157
def blob_dog(image, **kwargs): ...
158
class ORB(n_keypoints=500, **kwargs): ...
159
class SIFT(**kwargs): ...
160
```
161

162
[Feature Detection](./features.md)
163

164
### Image Measurement and Analysis
165

166
Measure geometric properties, analyze regions, compute image moments, and extract quantitative information from processed images.
167

168
```python { .api }
169
def label(image, background=None, **kwargs): ...
170
def regionprops(label_image, intensity_image=None, **kwargs): ...
171
def find_contours(image, level=0.5, **kwargs): ...
172
def moments(image, order=3): ...
173
def centroid(image): ...
174
def perimeter(image, neighbourhood=4): ...
175
```
176

177
[Measurement](./measurement.md)
178

179
### Image Input/Output
180

181
Read and write images in various formats with support for image collections, metadata handling, and multi-frame images.
182

183
```python { .api }
184
def imread(fname, **kwargs): ...
185
def imsave(fname, arr, **kwargs): ...
186
class ImageCollection(pattern, **kwargs): ...
187
def imread_collection(pattern, **kwargs): ...
188
```
189

190
[Input/Output](./io.md)
191

192
### Restoration and Denoising
193

194
Restore degraded images through deconvolution, denoising, and inpainting algorithms for image quality improvement.
195

196
```python { .api }
197
def denoise_tv_chambolle(image, weight=0.1, **kwargs): ...
198
def denoise_bilateral(image, **kwargs): ...
199
def denoise_wavelet(image, **kwargs): ...
200
def richardson_lucy(image, psf, num_iter=50, **kwargs): ...
201
def inpaint_biharmonic(image, mask, **kwargs): ...
202
```
203

204
[Restoration](./restoration.md)
205

206
### Intensity and Exposure Adjustment
207

208
Adjust image intensity, perform histogram operations, and enhance image contrast for better visualization and analysis.
209

210
```python { .api }
211
def rescale_intensity(image, in_range='image', out_range='dtype'): ...
212
def equalize_hist(image, nbins=256, mask=None): ...
213
def equalize_adapthist(image, kernel_size=None, **kwargs): ...
214
def adjust_gamma(image, gamma=1, gain=1): ...
215
def histogram(image, nbins=256, source_range='image', normalize=False): ...
216
```
217

218
[Exposure](./exposure.md)
219

220
### Drawing and Visualization
221

222
Create geometric shapes, draw on images, and generate synthetic test patterns for visualization and algorithm testing.
223

224
```python { .api }
225
def line(r0, c0, r1, c1): ...
226
def circle_perimeter(r, c, radius, method='bresenham', shape=None): ...
227
def polygon(r, c, shape=None): ...
228
def disk(center, radius, shape=None): ...
229
def rectangle(start, end=None, extent=None, shape=None): ...
230
def set_color(image, coords, color, alpha=1): ...
231
```
232

233
[Drawing](./drawing.md)
234

235
### Test Data and Examples
236

237
Access example images and synthetic datasets for testing algorithms and learning image processing techniques.
238

239
```python { .api }
240
def camera(): ...
241
def coins(): ...
242
def astronaut(): ...
243
def binary_blobs(length=512, blob_size_fraction=0.1, n_dim=2, **kwargs): ...
244
def checkerboard(): ...
245
```
246

247
[Data](./data.md)
248

249
### Utilities and Data Types
250

251
Convert between data types, generate noise, manipulate arrays, and access utility functions for image processing workflows.
252

253
```python { .api }
254
def img_as_float(image, force_copy=False): ...
255
def img_as_uint(image, force_copy=False): ...
256
def img_as_ubyte(image, force_copy=False): ...
257
def random_noise(image, mode='gaussian', **kwargs): ...
258
def crop(ar, crop_width, **kwargs): ...
259
def montage(arr_in, **kwargs): ...
260
```
261

262
[Utilities](./utilities.md)
263

264
## Types
265

266
```python { .api }
267
from typing import Union, Optional, Tuple, List, Dict, Any
268
from numpy.typing import NDArray
269
import numpy as np
270

271
# Core image type
272
Image = NDArray[np.number]
273

274
# Common parameter types
275
Sigma = Union[float, Tuple[float, ...]]
276
Footprint = Optional[NDArray[np.bool_]]
277
Shape = Tuple[int, ...]
278
ColorTuple = Tuple[float, float, float]
279
```