# Autocrop_kh
#### Automatic Document Segmentation and Cropping for Khmer IDs, Passport and Documents
Autocrop_kh is a Python package for automatic document segmentation and cropping, with a focus on Khmer IDs, Passport and other documents. It uses a DeepLabV3 model training on Khmer ID, Passport document datasets to accurately segment and extract documents from images.
License: [Apache-2.0 License](https://github.com/MetythornPenn/sdab/blob/main/LICENSE)
## Installation
#### Install from source
```sh
# clone repo
git clone https://github.com/MetythornPenn/autocrop_kh.git
# install lib from source
pip install -e .
```
#### Install from PyPI
```sh
pip install autocrop-kh
```
## Usage
#### Python Script
```python
import torch
import cv2
import requests
import os
from autocrop_kh import autocrop
# Function to download files from URLs
def download_file(url, local_filename):
with requests.get(url, stream=True) as r:
r.raise_for_status()
with open(local_filename, 'wb') as f:
for chunk in r.iter_content(chunk_size=8192):
f.write(chunk)
print(f"Downloaded: {local_filename}")
# URLs for the image and model
img_url = "https://github.com/MetythornPenn/autocrop_kh/raw/main/sample/img-1.jpg"
model_url = "https://github.com/MetythornPenn/autocrop_kh/raw/main/models/autocrop_model_v2.onnx"
# Local paths to save the files
img_path = "img-1.jpg"
model_path = "autocrop_model_v2.onnx"
# Download the image and model files
download_file(img_url, img_path)
download_file(model_url, model_path)
# Verify the files are correctly downloaded
if not os.path.exists(img_path):
raise FileNotFoundError(f"Image file {img_path} was not found.")
if not os.path.exists(model_path):
raise FileNotFoundError(f"Model file {model_path} was not found.")
# Specify device (CPU or CUDA or Apple Silicon GPU)
if torch.cuda.is_available():
device = "cuda" # Use NVIDIA GPU (if available)
elif torch.backends.mps.is_available():
device = "mps" # Use Apple Silicon GPU (if available)
else:
device = "cpu" # Default to CPU if no GPU is available
# Perform document extraction
extracted_document = autocrop(img_path=img_path, model_path=model_path, device=device)
# Save the extracted document
output_path = "extracted_document.jpg"
cv2.imwrite(output_path, extracted_document[:, :, ::-1]) # Convert back to BGR for saving
print(f"Extracted document saved to {output_path}")
```
- `img_path`: Path of the input image file.
- `model_path`: Path to the pre-trained model (local path and support both .onnx and .pth).
- `device`: Specify `cpu` or `cuda` or `mps` (default is `gpu`).
- `output_path`: Path where the extracted document image will be saved.
#### Result:
<p align="center">
<img src="sample/img-1.jpg" alt="Left Image" width="45%">
<img src="sample/result-img-1.png" alt="Right Image" width="45%">
</p>
<p align="center">
<img src="sample/img-5.png" alt="Left Image" width="45%">
<img src="sample/result-img-5.png" alt="Right Image" width="45%">
</p>
#### Running as API & Web
```sh
# clone repo
git clone https://github.com/MetythornPenn/autocrop_kh.git
# go to directory
cd autocrop
# install libraries
pip3 install -r requirements.txt
# start server (http://localhost:5555/docs)
make server
# start client ((http://localhost:7860))
make client
```
**Noted** : This model was trained with 25000 datasets include opensource data and my custom synthetic data.
## Reference
- Inspired by [DeepLabV3](https://paperswithcode.com/method/deeplabv3)
- [Publish python package to PyPI](https://www.youtube.com/watch?v=90PWQEc--6k)
Raw data
{
"_id": null,
"home_page": "https://github.com/MetythornPenn/autocrop_kh.git",
"name": "autocrop-kh",
"maintainer": null,
"docs_url": null,
"requires_python": null,
"maintainer_email": null,
"keywords": "autocrop_kh",
"author": "Metythorn Penn",
"author_email": "metythorn@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/90/26/38631b9f863a44818d0c6d03502967d51991d89ad673febf505b3c6e6f09/autocrop_kh-0.0.3.tar.gz",
"platform": null,
"description": "# Autocrop_kh\n\n#### Automatic Document Segmentation and Cropping for Khmer IDs, Passport and Documents\n\nAutocrop_kh is a Python package for automatic document segmentation and cropping, with a focus on Khmer IDs, Passport and other documents. It uses a DeepLabV3 model training on Khmer ID, Passport document datasets to accurately segment and extract documents from images.\n\nLicense: [Apache-2.0 License](https://github.com/MetythornPenn/sdab/blob/main/LICENSE)\n\n## Installation\n\n#### Install from source\n\n```sh\n\n# clone repo \ngit clone https://github.com/MetythornPenn/autocrop_kh.git\n\n# install lib from source\npip install -e .\n\n```\n\n#### Install from PyPI\n```sh\npip install autocrop-kh\n```\n\n## Usage\n\n#### Python Script\n\n```python\nimport torch\nimport cv2\nimport requests\nimport os\nfrom autocrop_kh import autocrop\n\n# Function to download files from URLs\ndef download_file(url, local_filename):\n with requests.get(url, stream=True) as r:\n r.raise_for_status()\n with open(local_filename, 'wb') as f:\n for chunk in r.iter_content(chunk_size=8192):\n f.write(chunk)\n print(f\"Downloaded: {local_filename}\")\n\n# URLs for the image and model\nimg_url = \"https://github.com/MetythornPenn/autocrop_kh/raw/main/sample/img-1.jpg\"\nmodel_url = \"https://github.com/MetythornPenn/autocrop_kh/raw/main/models/autocrop_model_v2.onnx\"\n\n# Local paths to save the files\nimg_path = \"img-1.jpg\"\nmodel_path = \"autocrop_model_v2.onnx\"\n\n# Download the image and model files\ndownload_file(img_url, img_path)\ndownload_file(model_url, model_path)\n\n# Verify the files are correctly downloaded\nif not os.path.exists(img_path):\n raise FileNotFoundError(f\"Image file {img_path} was not found.\")\nif not os.path.exists(model_path):\n raise FileNotFoundError(f\"Model file {model_path} was not found.\")\n\n# Specify device (CPU or CUDA or Apple Silicon GPU)\nif torch.cuda.is_available():\n device = \"cuda\" # Use NVIDIA GPU (if available)\nelif torch.backends.mps.is_available():\n device = \"mps\" # Use Apple Silicon GPU (if available)\nelse:\n device = \"cpu\" # Default to CPU if no GPU is available\n\n# Perform document extraction\nextracted_document = autocrop(img_path=img_path, model_path=model_path, device=device)\n\n# Save the extracted document\noutput_path = \"extracted_document.jpg\"\ncv2.imwrite(output_path, extracted_document[:, :, ::-1]) # Convert back to BGR for saving\n\nprint(f\"Extracted document saved to {output_path}\")\n\n\n```\n\n- `img_path`: Path of the input image file.\n- `model_path`: Path to the pre-trained model (local path and support both .onnx and .pth).\n- `device`: Specify `cpu` or `cuda` or `mps` (default is `gpu`).\n- `output_path`: Path where the extracted document image will be saved.\n\n#### Result:\n\n<p align=\"center\">\n <img src=\"sample/img-1.jpg\" alt=\"Left Image\" width=\"45%\">\n <img src=\"sample/result-img-1.png\" alt=\"Right Image\" width=\"45%\">\n</p>\n\n<p align=\"center\">\n <img src=\"sample/img-5.png\" alt=\"Left Image\" width=\"45%\">\n <img src=\"sample/result-img-5.png\" alt=\"Right Image\" width=\"45%\">\n</p>\n\n\n#### Running as API & Web\n```sh\n# clone repo\ngit clone https://github.com/MetythornPenn/autocrop_kh.git\n\n# go to directory\ncd autocrop\n\n# install libraries\npip3 install -r requirements.txt\n\n# start server (http://localhost:5555/docs)\nmake server\n\n# start client ((http://localhost:7860))\nmake client \n\n```\n**Noted** : This model was trained with 25000 datasets include opensource data and my custom synthetic data.\n## Reference \n- Inspired by [DeepLabV3](https://paperswithcode.com/method/deeplabv3)\n- [Publish python package to PyPI](https://www.youtube.com/watch?v=90PWQEc--6k)\n\n",
"bugtrack_url": null,
"license": "Apache Software License 2.0",
"summary": "Document Extraction Inference API using DeepLabV3 with Pretrain Model",
"version": "0.0.3",
"project_urls": {
"Homepage": "https://github.com/MetythornPenn/autocrop_kh.git"
},
"split_keywords": [
"autocrop_kh"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "e353fe1f5a524f3f5fb788b010ef442a7b7cd3b516320b7df0eedc8bdded6a52",
"md5": "1bdbb32e996e19c4495e5c51f897e70e",
"sha256": "eb4f35b9a6c5272db8d4bb5ee2f80ac342228d0cb9f29a1b4a2aa2bc2a8e865a"
},
"downloads": -1,
"filename": "autocrop_kh-0.0.3-py3-none-any.whl",
"has_sig": false,
"md5_digest": "1bdbb32e996e19c4495e5c51f897e70e",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 9514,
"upload_time": "2024-10-24T09:16:29",
"upload_time_iso_8601": "2024-10-24T09:16:29.027190Z",
"url": "https://files.pythonhosted.org/packages/e3/53/fe1f5a524f3f5fb788b010ef442a7b7cd3b516320b7df0eedc8bdded6a52/autocrop_kh-0.0.3-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "902638631b9f863a44818d0c6d03502967d51991d89ad673febf505b3c6e6f09",
"md5": "b24eab759f1a5d7e1c484d15ce4ea777",
"sha256": "4af7de8c7edd9c2a8420dcc1e9bae1d899c6e8158d06b2bc3edfd13e7dc20607"
},
"downloads": -1,
"filename": "autocrop_kh-0.0.3.tar.gz",
"has_sig": false,
"md5_digest": "b24eab759f1a5d7e1c484d15ce4ea777",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 9505,
"upload_time": "2024-10-24T09:16:30",
"upload_time_iso_8601": "2024-10-24T09:16:30.386105Z",
"url": "https://files.pythonhosted.org/packages/90/26/38631b9f863a44818d0c6d03502967d51991d89ad673febf505b3c6e6f09/autocrop_kh-0.0.3.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-10-24 09:16:30",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "MetythornPenn",
"github_project": "autocrop_kh",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"requirements": [
{
"name": "torch",
"specs": []
},
{
"name": "torchvision",
"specs": []
},
{
"name": "numpy",
"specs": []
},
{
"name": "opencv-python-headless",
"specs": []
},
{
"name": "fastapi",
"specs": []
},
{
"name": "request",
"specs": []
},
{
"name": "certifi",
"specs": []
}
],
"lcname": "autocrop-kh"
}
JSON
