# Lightning-fast image color clustering with C-based RGB localization/euclidean distance calculation. Supports DBSCAN/HDBSCAN, Shapely geometry.
## pip install locatecolorcluster
### Tested against Python 3.11 / Windows 10
## Important: A C/C++ compiler is necessary!
## Advantages:
### Flexibility:
The module supports both DBSCAN and HDBSCAN clustering algorithms, allowing users to choose the method that best fits their needs.
### Parallelization:
The code leverages parallelization capabilities in clustering algorithms, specifically in DBSCAN, HDBSCAN and cythoneuclideandistance, by allowing users to specify the number of parallel jobs.
### Speed Optimization:
The module exhibits exceptional speed, outperforming standard libraries like SciPy in Euclidean distance calculations by a factor of four. This performance boost is crucial for large-scale image processing tasks.
### RGB Localization with C:
The module includes efficient C implementations for RGB localization (300x faster than Python - search_colors method), enhancing the speed and accuracy of color identification in images.
### Visualization:
The module includes methods (draw_results and draw_results_real_size) for visualizing the clustered results, making it easier for users to interpret and analyze the outcomes.
### Shapely Geometry:
The module provides Shapely geometry information for each cluster, offering additional insights into the spatial characteristics of the clusters.
```python
Dependencies:
- Cython and a C/C++ compiler!
- os.path
- sys
- cv2
- scipy.spatial.distance.pdist, squareform
- cythoneuclideandistance
- a_cv_imwrite_imread_plus
- locate_pixelcolor_c
- numpy
- shapely.geometry
- sklearn.cluster.DBSCAN, HDBSCAN
- numexpr
- a_cv2_easy_resize
Usage:
- Create an instance of ColorCluster by providing an image path, optional parameters for resizing,
and interpolation method.
- Use the find_colors method to search for specific colors in the image.
- Calculate the Euclidean distance matrix using the calculate_euclidean_matrix method with the desired backend.
- Apply clustering algorithms (DBSCAN or HDBSCAN) using the get_dbscan_labels or get_hdbscan_labels methods.
- Extract clusters and visualize results using get_clusters, draw_results, and draw_results_real_size methods.
- Obtain Shapely geometry information for each cluster using get_shapely method.
Example:
import cv2
from locatecolorcluster import ColorCluster, get_range_of_colors
# Some valid color inputs
colors = get_range_of_colors(
start=(0, 0, 0),
end=(10, 10, 10),
rgb_or_bgr="bgr",
min_r=0,
max_r=10,
min_g=0,
max_g=10,
min_b=0,
max_b=10,
)
colors = ((0, 0, 0),)
colors = get_range_of_colors(
start=(100, 0, 0),
end=(255, 0, 0),
rgb_or_bgr="bgr",
min_r=100,
max_r=255,
min_g=0,
max_g=10,
min_b=0,
max_b=10,
)
cbackend = (
ColorCluster(
img=r'https://avatars.githubusercontent.com/u/77182807?s=400&u=b3398787384abf38d62c6f080195550df64f3990&v=4',
max_width=200,
max_height=200,
interpolation=cv2.INTER_NEAREST,
)
.find_colors(colors=colors, reverse_colors=False)
.calculate_euclidean_matrix(backend="C", memorylimit_mb=10000) # Define a memory limit for the C backend - It's 4x faster than scipy, but if the array is too big for the memory, the process crashes with 0xc0000005
.get_dbscan_labels(eps=3, min_samples=10, algorithm="auto", leaf_size=30, n_jobs=5)
.get_clusters()
.draw_results(folder=r"C:\myimageresults\1", color=(255, 0, 255))
.draw_results_real_size(folder=r"C:\myimageresults\2", color=(255, 255, 0))
.get_shapely()
)
shapelydata = cbackend.shapelydata
print(shapelydata)
```
Raw data
{
"_id": null,
"home_page": "https://github.com/hansalemaos/locatecolorcluster",
"name": "locatecolorcluster",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "DBSCAN,HDBSCAN,euclidean",
"author": "Johannes Fischer",
"author_email": "aulasparticularesdealemaosp@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/72/2a/2507c978a79038ec9ccb5ed83bf72a47ac13badeefd5563bdb60ac8c6c2a/locatecolorcluster-0.10.tar.gz",
"platform": null,
"description": "\r\n# Lightning-fast image color clustering with C-based RGB localization/euclidean distance calculation. Supports DBSCAN/HDBSCAN, Shapely geometry. \r\n\r\n## pip install locatecolorcluster\r\n\r\n### Tested against Python 3.11 / Windows 10\r\n\r\n## Important: A C/\u0003C++ compiler is necessary!\r\n\r\n## Advantages:\r\n\r\n\r\n### Flexibility:\r\n\r\nThe module supports both DBSCAN and HDBSCAN clustering algorithms, allowing users to choose the method that best fits their needs.\r\n\r\n\r\n### Parallelization:\r\n\r\nThe code leverages parallelization capabilities in clustering algorithms, specifically in DBSCAN, HDBSCAN and cythoneuclideandistance, by allowing users to specify the number of parallel jobs.\r\n\r\n\r\n### Speed Optimization:\r\n\r\nThe module exhibits exceptional speed, outperforming standard libraries like SciPy in Euclidean distance calculations by a factor of four. This performance boost is crucial for large-scale image processing tasks.\r\n\r\n\r\n### RGB Localization with C:\r\n\r\nThe module includes efficient C implementations for RGB localization (300x faster than Python - search_colors method), enhancing the speed and accuracy of color identification in images.\r\n\r\n\r\n### Visualization:\r\n\r\nThe module includes methods (draw_results and draw_results_real_size) for visualizing the clustered results, making it easier for users to interpret and analyze the outcomes.\r\n\r\n\r\n### Shapely Geometry:\r\n\r\nThe module provides Shapely geometry information for each cluster, offering additional insights into the spatial characteristics of the clusters.\r\n\r\n\r\n\r\n\r\n\r\n```python\r\nDependencies:\r\n\t- Cython and a C/C++ compiler!\r\n\t- os.path\r\n\t- sys\r\n\t- cv2\r\n\t- scipy.spatial.distance.pdist, squareform\r\n\t- cythoneuclideandistance\r\n\t- a_cv_imwrite_imread_plus\r\n\t- locate_pixelcolor_c\r\n\t- numpy\r\n\t- shapely.geometry\r\n\t- sklearn.cluster.DBSCAN, HDBSCAN\r\n\t- numexpr\r\n\t- a_cv2_easy_resize\r\n\r\nUsage:\r\n\t- Create an instance of ColorCluster by providing an image path, optional parameters for resizing,\r\n\t and interpolation method.\r\n\t- Use the find_colors method to search for specific colors in the image.\r\n\t- Calculate the Euclidean distance matrix using the calculate_euclidean_matrix method with the desired backend.\r\n\t- Apply clustering algorithms (DBSCAN or HDBSCAN) using the get_dbscan_labels or get_hdbscan_labels methods.\r\n\t- Extract clusters and visualize results using get_clusters, draw_results, and draw_results_real_size methods.\r\n\t- Obtain Shapely geometry information for each cluster using get_shapely method.\r\n\r\nExample:\r\n\r\n\timport cv2\r\n\r\n\tfrom locatecolorcluster import ColorCluster, get_range_of_colors\r\n\t# Some valid color inputs\r\n\tcolors = get_range_of_colors(\r\n\t\tstart=(0, 0, 0),\r\n\t\tend=(10, 10, 10),\r\n\t\trgb_or_bgr=\"bgr\",\r\n\t\tmin_r=0,\r\n\t\tmax_r=10,\r\n\t\tmin_g=0,\r\n\t\tmax_g=10,\r\n\t\tmin_b=0,\r\n\t\tmax_b=10,\r\n\t)\r\n\tcolors = ((0, 0, 0),)\r\n\tcolors = get_range_of_colors(\r\n\t\tstart=(100, 0, 0),\r\n\t\tend=(255, 0, 0),\r\n\t\trgb_or_bgr=\"bgr\",\r\n\t\tmin_r=100,\r\n\t\tmax_r=255,\r\n\t\tmin_g=0,\r\n\t\tmax_g=10,\r\n\t\tmin_b=0,\r\n\t\tmax_b=10,\r\n\t)\r\n\tcbackend = (\r\n\t\tColorCluster(\r\n\t\t\timg=r'https://avatars.githubusercontent.com/u/77182807?s=400&u=b3398787384abf38d62c6f080195550df64f3990&v=4',\r\n\t\t\tmax_width=200,\r\n\t\t\tmax_height=200,\r\n\t\t\tinterpolation=cv2.INTER_NEAREST,\r\n\t\t)\r\n\t\t.find_colors(colors=colors, reverse_colors=False)\r\n\t\t.calculate_euclidean_matrix(backend=\"C\", memorylimit_mb=10000) # Define a memory limit for the C backend - It's 4x faster than scipy, but if the array is too big for the memory, the process crashes with 0xc0000005\r\n\t\t.get_dbscan_labels(eps=3, min_samples=10, algorithm=\"auto\", leaf_size=30, n_jobs=5)\r\n\t\t.get_clusters()\r\n\t\t.draw_results(folder=r\"C:\\myimageresults\\1\", color=(255, 0, 255))\r\n\t\t.draw_results_real_size(folder=r\"C:\\myimageresults\\2\", color=(255, 255, 0))\r\n\t\t.get_shapely()\r\n\t)\r\n\tshapelydata = cbackend.shapelydata\r\n\tprint(shapelydata)\r\n\r\n\r\n\r\n```\r\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "Lightning-fast image color clustering with C-based RGB localization/euclidean distance calculation. Supports DBSCAN/HDBSCAN, Shapely geometry.",
"version": "0.10",
"project_urls": {
"Homepage": "https://github.com/hansalemaos/locatecolorcluster"
},
"split_keywords": [
"dbscan",
"hdbscan",
"euclidean"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "866255c18ab7db7c136ab84c12d8d6b5a0c5ecac111dd89423d5e07225407325",
"md5": "4872efe395f09d91a796537f9f6c07a6",
"sha256": "5f390b32aa01b91f46116244aa8c5187665c07c77937f434f352f72c1a841071"
},
"downloads": -1,
"filename": "locatecolorcluster-0.10-py3-none-any.whl",
"has_sig": false,
"md5_digest": "4872efe395f09d91a796537f9f6c07a6",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 76782,
"upload_time": "2023-11-14T04:43:21",
"upload_time_iso_8601": "2023-11-14T04:43:21.484887Z",
"url": "https://files.pythonhosted.org/packages/86/62/55c18ab7db7c136ab84c12d8d6b5a0c5ecac111dd89423d5e07225407325/locatecolorcluster-0.10-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "722a2507c978a79038ec9ccb5ed83bf72a47ac13badeefd5563bdb60ac8c6c2a",
"md5": "08ef7efe5cc3568762abcfe91ccd641c",
"sha256": "9923089dd8accb1377e2e3e1b01e5df4f6d6bc85412a900da44fc25599730071"
},
"downloads": -1,
"filename": "locatecolorcluster-0.10.tar.gz",
"has_sig": false,
"md5_digest": "08ef7efe5cc3568762abcfe91ccd641c",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 75854,
"upload_time": "2023-11-14T04:43:23",
"upload_time_iso_8601": "2023-11-14T04:43:23.810205Z",
"url": "https://files.pythonhosted.org/packages/72/2a/2507c978a79038ec9ccb5ed83bf72a47ac13badeefd5563bdb60ac8c6c2a/locatecolorcluster-0.10.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-11-14 04:43:23",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "hansalemaos",
"github_project": "locatecolorcluster",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"requirements": [],
"lcname": "locatecolorcluster"
}
JSON
