# dkimageapp

**Repository Path**: weihaitong/dkimageapp

## Basic Information

- **Project Name**: dkimageapp
- **Description**: dkimageapp是61Duke于2020年整合工作与学习研发关于图像处理的实用工具Python依赖库模块，该模块统一规范的API接口，承载个人未来工作中开发有关图像处理实际应用。该模块由算法科学家61Duke独家研发，提供图像裁剪、合并、图像格式转换、色调占比提取、颜色相似度、图像定位和图像检索等功能。dkimageapp模块为各个项目和业务提供支持。

- **Primary Language**: Python
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2020-12-30
- **Last Updated**: 2022-06-08

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

==========
dkimageapp
==========


.. image:: https://img.shields.io/pypi/v/dkimageapp.svg
        :target: https://pypi.python.org/pypi/dkimageapp

.. image:: https://img.shields.io/travis/61Duke/dkimageapp.svg
        :target: https://travis-ci.org/61Duke/dkimageapp

.. image:: https://readthedocs.org/projects/dkimageapp/badge/?version=latest
        :target: https://dkimageapp.readthedocs.io/en/latest/?badge=latest
        :alt: Documentation Status



dkimageapp是61Duke于2020年整合工作与学习研发关于图像处理的实用工具Python依赖库模块，该模块统一规范的API接口，
承载个人未来工作中开发有关图像处理实际应用。该模块由算法科学家61Duke独家研发，提供图像裁剪、合并、图像格式转换、
色调占比提取、颜色相似度、图像定位和图像检索等功能。dkimageapp模块为各个项目和业务提供支持。


* Free software: MIT license


文档
--------
导入dkimageapp库::

    import dkimageapp

图像合并
~~~~~~~~~~~
图像上加入图标和文字::

    dkimageapp.core.addition_icon_autograph(
        source_image_path="imgs/IMG_2229.JPG",
        output_path="source_output.png",
        icon_path="imgs/icon.jpg",
        autograph="@61Duke",
        autograph_method="right_down",
        icon_method="right_down_circle"
    )


图像色彩
~~~~~~~~~~~
提取图像主色调::

    dominant = dkimageapp.core.colour_dominant(image_path="imgs/IMG_2229.JPG")
    print(dominant)
    # (0.51, (35, 18, 26))

提取图像色调占比::

    palette = dkimageapp.core.colour_ratio(image_path="imgs/IMG_2229.JPG")
    print(palette)
    # [(0.51, (35, 18, 26)), (0.19, (121, 41, 59)), (0.09, (142, 129, 110)), (0.07, (209, 129, 118)), (0.05, (214, 35, 62)), (0.04, (131, 102, 77)), (0.02, (225, 42, 115)), (0.02, (207, 188, 148))]

提取颜色与颜色的相似度::

    similarity = dkimageapp.core.colour_similarity(base_rgb=(255, 0, 0), compare_rgb=(255, 5, 50))
    print(similarity)
    # 0.4480129357712845

图像定位
~~~~~~~~~~~
图像GPS信息::

    gps = dkimageapp.core.location_gps(image_path="imgs/IMG_2229.JPG")
    print(gps)
    """
    {
        "GPS": {
            "GPSLatitudeRef": "N",
            "GPSLatitude": 18.292461111111113,
            "GPSLongitudeRef": "E",
            "GPSLongitude": 109.52616944444445,
            "GPSAltitudeRef": "0",
            "GPSAltitude": "41477/8571",
            "GPSAddress": "海南省三亚市吉阳区河东区街道迎宾路329号三亚千古情景区"
        },
        "date": "2019:12:20 14:29:37"
    }
    """

图像裁剪
~~~~~~~~~~~
视频分割成音频、视频添加音频功能依靠 ffmpeg。

FFmpeg是一套可以用来记录、转换数字音频、视频，并能将其转化为流的开源计算机程序。采用LGPL或GPL许可证。它提供了录制、转换以及流化音视频的完整解决方案。它包含了非常先进的音频/视频编解码库libavcodec，为了保证高可移植性和编解码质量，libavcodec里很多code都是从头开发的。
下载有命令行可以使用。

将GIF动图切割成一帧一帧图像::

    dkimageapp.core.crop_gif2images(
        source_image_path="imgs/test.gif",
        output_folder="source_output"
    )

图像坐标部分裁剪::

    dkimageapp.core.crop_part_image(
    source_image_path="imgs/IMG_2229.JPG",
    output_path="source_output.png",
    box=(300, 200, 700, 800)
    )

图像拼接
~~~~~~~~~~~
将一帧一帧图像和成GIF动图::

    dkimageapp.core.stitch_images2gif(
        source_images_path=[
            "imgs/20191017185607410.png",
            "imgs/20191017185634484.png",
            "imgs/icon.jpg",
            "imgs/IMG_2229.JPG",
            "imgs/IMG_2994.JPG"
        ],
        output_path="source_output.gif"
    )

图像按照指定规格拼接::

    dkimageapp.core.stitch_images(
        source_images_path=[
            "imgs/20191017185607410.png",
            "imgs/20191017185634484.png",
            "imgs/icon.jpg",
            "imgs/IMG_2229.JPG",
            "imgs/IMG_2994.JPG"
        ],
        output_path="output.png",
        layout=(4, 2),
        each_image_shape=(0, 0)
    )

图像检索
~~~~~~~~~~~
自动编码机::

    nearest_images = dkimageapp.core.retrieval_images_by_autoencoder(
        fit_images_path=[
            "imgs/20191017185607410.png",
            "imgs/20191017185634484.png",
            "imgs/icon.jpg",
            "imgs/IMG_2229.JPG",
            "imgs/IMG_2994.JPG"
        ],
        target_images_path=["122.JPG", "017.JPG"],
        scale=28,
        epochs=15,
        batch_size=32,
        n_neigh=5,
        shuffle=True
    )
    print(nearest_images)
    """
    >>> {
            ...    "122.JPG": [
            ...        (1.0, "data/124.JPG"),
            ...        (0.021841750088447576, "data/535.JPG"),
            ...        (0.013354248283854897, "data/605.JPG"),
            ...        (0.009994627751588991, "data/402.JPG"),
            ...        (0.0, "data/134.JPG")
            ...    ],
            ...    "017.JPG": [
            ...        (0.11133096949872578, "data/510.JPG"),
            ...        (0.03817764172216209, "data/471.JPG"),
            ...        (0.02644454978659949, "data/IMG_2044.JPG"),
            ...        (0.001042320143904596, "data/130.JPG"),
            ...        (0.0, "data/472.JPG")
            ...    ]
    >>> }

    """

资源
------
制作微信国旗头像，别再@微信官方

自动编码器（卷积神经网络CNN）构建一个简单的图像检索系统：

* https://github.com/nathanhubens/Unsupervised-Image-Retrieval
* https://towardsdatascience.com/build-a-simple-image-retrieval-system-with-an-autoencoder-673a262b7921


Credits
-------

This package was created with Cookiecutter_ and the `audreyr/cookiecutter-pypackage`_ project template.

.. _Cookiecutter: https://github.com/audreyr/cookiecutter
.. _`audreyr/cookiecutter-pypackage`: https://github.com/audreyr/cookiecutter-pypackage