{
    "info": {
        "author": "",
        "author_email": "",
        "bugtrack_url": null,
        "classifiers": [
            "License :: OSI Approved :: MIT License",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 3"
        ],
        "description": "# Facial Recognition and Alignment\n## What's this?\n\nThis repository contains a refactored implementation of David Sandberg's [FaceNet](https://github.com/davidsandberg/facenet) and [InsightFace](https://github.com/deepinsight/insightface) for facial recognition. It also contains an implementation of [MTCNN](https://github.com/ipazc/mtcnn) and [Faceboxes](https://github.com/TropComplique/FaceBoxes-tensorflow) for face cropping and alignment. What is in the refactor:\n\n- Made algorithms easily and efficiently usable with [convenience classes](https://github.com/armanrahman22/facenet/tree/master/facenet_sandberg/inference). \n- Added much more efficient methods of batch processing face recognition and alignment\n- Added true face alignment (with affine transformation) to align face to bottom-center of image: [code](https://github.com/armanrahman22/facenet/blob/f6cb32a193925002da41fb491c52bb85384bee55/facenet_sandberg/utils.py#L187)\n- Added proportional margin to alignment as per this [issue](https://github.com/davidsandberg/facenet/issues/283)\n- Ability to easily switch between [insightface](https://github.com/armanrahman22/facenet/blob/master/facenet_sandberg/inference/insightface_encoder.py) and [facenet](https://github.com/armanrahman22/facenet/blob/master/facenet_sandberg/inference/facenet_encoder.py) at [inference time](https://github.com/armanrahman22/facenet/blob/master/facenet_sandberg/inference/identifier.py)\n\nMore information on customizing and implementing new face detection algorithms can be found [here](./algorithms/README.md).\n\n## Installation\nTo use in other projects, this implementation can be pip installed as follows:\n```\npip install facenet_sandberg\n```\n\nTo use locally:\n1. Clone repo\n2. cd to base directory with setup.py \n3. run:\n```\npip install -e .\n```\n^(installs package in [development mode](https://setuptools.readthedocs.io/en/latest/setuptools.html#development-mode))\n\n## Important Requirements\n1. Python 3.5 \n2. Tensorflow==1.7\n3. Tensorlayer==1.7\nThe rest is specified in [requirements.txt](https://github.com/armanrahman22/facenet/blob/master/requirements.txt)\n\n## Models\nLinks to pretrained models: \n\n- [Facenet](https://redcrossstorage.blob.core.windows.net/images/facenet_model.pb)\n  - Uses RGB images of size 160x160\n- [Insightface.zip](https://redcrossstorage.blob.core.windows.net/images/insightface_ckpt.zip)\n  - Uses BGR images of size 112x112\n\n## Datasets\nLinks to download training datasets (!big files!):\n\n- [Emore](https://redcrossstorage.blob.core.windows.net/datasets/faces_emore.zip) \n- [MSM_refined_112x112](https://redcrossstorage.blob.core.windows.net/datasets/faces_ms1m-refine-v2_112x112.zip)\n- [VGG2_112x112](https://redcrossstorage.blob.core.windows.net/datasets/faces_vgg2_112x112.zip)\n\n## Image directory structure\nThis repo assumes images are in [LFW format](http://vis-www.cs.umass.edu/lfw/README.txt):\n```\n-/base_images_folder\n  -/person_1\n    -person_1_0001.jpg\n    -person_1_0002.jpg\n    -person_1_0003.jpg\n  -/person_2\n    -person_2_0001.jpg\n    -person_2_0002.jpg\n  ...\n```\n\nIf your dataset is not like this you can use [lfw.py](https://github.com/armanrahman22/facenet/blob/master/facenet_sandberg/lfw.py) to put your images into the right format like so (from facenet_sandberg/facenet_sandberg):\n```\npython lfw.py --image_directory PATH_TO_YOUR_BASE_IMAGE_DIRECTORY\n```\n\n## Alignment\nAlignment is done with a combination of Faceboxes and MTCNN. While Faceboxes is more accurate and works with more images than MTCNN, it does not return [facial landmarks](https://raw.githubusercontent.com/ipazc/mtcnn/master/result.jpg). Whichever algorithm returns more results is used.\n\nUse the [align_dataset.py](https://github.com/armanrahman22/facenet/blob/master/facenet_sandberg/align_dataset.py) script to align an entire image directory:\n```\npython align_dataset.py --input_dir PATH_TO_YOUR_BASE_IMAGE_DIRECTORY \\\n                        --output_dir PATH_TO_OUTPUT_ALIGNED_IMAGES \\\n                        --facenet_model_checkpoint PATH_TO_PRETRAINED_FACENET_MODEL \\\n                        --image_height DESIRED_IMAGE_HEIGHT \\\n                        --image_width DESIRED_IMAGE_WIDTH \\\n                        --margin DESIRED_PROPORTIONAL_MARGIN \\\n                        --scale_factor DESIRED_SCALE_FACTOR \\\n                        --steps_threshold DESIRED_STEPS \\\n                        --detect_multiple_faces \\\n                        --use_faceboxes \\\n                        --use_affine \\\n                        --num_processes NUM_PROCESSES_TO_USE\n```\n* Default values for most arguments are provided [here](https://github.com/armanrahman22/facenet/blob/f6cb32a193925002da41fb491c52bb85384bee55/facenet_sandberg/align_dataset.py#L262) \n\n## Generate Pairs.txt\nA pairs.txt file is used in training and testing. It follows this [format](http://vis-www.cs.umass.edu/lfw/README.txt). In order to generate your own pairs.txt run:\n```\npython align_dataset.py --image_dir PATH_TO_YOUR_BASE_IMAGE_DIRECTORY \\\n                        --pairs_file_name OUTPUT_NAME_OF_PAIRS_FILE \\\n                        --num_folds NUMBER_OF_FOLDS_FOR_CROSS_VALIDATION \\\n                        --num_matches_mismatches NUMBER_OF_MATCHES_AND_MISMATCHES\n```\n\n## Copyright\nMIT License from original repo https://github.com/davidsandberg/facenet/blob/master/LICENSE.md",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/armanrahman22/facenet",
        "keywords": "",
        "license": "MIT",
        "maintainer": "Arman Rahman",
        "maintainer_email": "armanrahman22@gmail.com",
        "name": "facenet-sandberg",
        "package_url": "https://pypi.org/project/facenet-sandberg/",
        "platform": "",
        "project_url": "https://pypi.org/project/facenet-sandberg/",
        "project_urls": {
            "Homepage": "https://github.com/armanrahman22/facenet"
        },
        "release_url": "https://pypi.org/project/facenet-sandberg/0.0.8/",
        "requires_dist": null,
        "requires_python": ">=3.4",
        "summary": "Face recognition using TensorFlow",
        "version": "0.0.8"
    },
    "last_serial": 4470049,
    "releases": {
        "0.0.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "ea43a58af990f318cc644f834d0dc65b",
                    "sha256": "176c614a5a0504bf9830dcaa69168d171e3dcb471987b947fad1b30b98d0477d"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.1-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "ea43a58af990f318cc644f834d0dc65b",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": null,
                "size": 2058319,
                "upload_time": "2018-11-05T08:23:00",
                "url": "https://files.pythonhosted.org/packages/57/66/7ac8ba6ed3ca1ec37ecd24a9c9fbae0a2c10ff6c8923dfd5042dc31a6ced/facenet_sandberg-0.0.1-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "5bf8dbd97b990c4d51f556d184cc8b17",
                    "sha256": "a8f563aa720a20b582fb847bcb80bd84f29396e7eb23f67ee5c133424e5e928b"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.1.tar.gz",
                "has_sig": false,
                "md5_digest": "5bf8dbd97b990c4d51f556d184cc8b17",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 66159,
                "upload_time": "2018-11-05T08:23:02",
                "url": "https://files.pythonhosted.org/packages/90/2a/f98454d085a46f1e969d1a10c8ceb20ae600f5e15ea78a6fdb84f16ecda4/facenet_sandberg-0.0.1.tar.gz"
            }
        ],
        "0.0.2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "e7f162145f6fa971cb3e631361be60ed",
                    "sha256": "009eb795f095ba186103990ee85e0e789e3dcc37799faae9c26685eccd4548ae"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.2.tar.gz",
                "has_sig": false,
                "md5_digest": "e7f162145f6fa971cb3e631361be60ed",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 74179,
                "upload_time": "2018-11-07T16:34:06",
                "url": "https://files.pythonhosted.org/packages/b4/83/fa1aa4383e34716d4f44192a502291057413241dd3c3c0b5b1234edea2b9/facenet_sandberg-0.0.2.tar.gz"
            }
        ],
        "0.0.3": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "128318b970c458e9d5ac84ec774943e2",
                    "sha256": "e40d2591ecaba7a65c5bd57a6c609ab9153ba258eb81a1c0b2e7939c4850b1ee"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.3.tar.gz",
                "has_sig": false,
                "md5_digest": "128318b970c458e9d5ac84ec774943e2",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 74215,
                "upload_time": "2018-11-07T16:56:09",
                "url": "https://files.pythonhosted.org/packages/fb/d7/80805a27d10baab304860576961725614418130933c04ab61e2fb8afc280/facenet_sandberg-0.0.3.tar.gz"
            }
        ],
        "0.0.4": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "c0f6e1e36d63d2acd3c876091df29374",
                    "sha256": "f65e75c5cc86c0842a0afe050e5a8f09b0722bc4d789344fb646027abd57a5c8"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.4.tar.gz",
                "has_sig": false,
                "md5_digest": "c0f6e1e36d63d2acd3c876091df29374",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 74495,
                "upload_time": "2018-11-08T10:08:12",
                "url": "https://files.pythonhosted.org/packages/63/a8/73a7d912dd6c0e84e56aa77045d3db634f54b205ab77b9d8f62bdbaac6b7/facenet_sandberg-0.0.4.tar.gz"
            }
        ],
        "0.0.5": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "37f20359ea80fc8bd84233f6cb0d7589",
                    "sha256": "51b3c5fce08ae539291c4ce6bc21dd9debfcd6c5b2794ec837a26c74ac4e7181"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.5.tar.gz",
                "has_sig": false,
                "md5_digest": "37f20359ea80fc8bd84233f6cb0d7589",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4",
                "size": 33784,
                "upload_time": "2018-11-09T12:09:57",
                "url": "https://files.pythonhosted.org/packages/6c/8b/b2f364a236260e70e48a07de93e7da51707bff6ee84c27c8ccadeae984f0/facenet_sandberg-0.0.5.tar.gz"
            }
        ],
        "0.0.6": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "246a0f24513aa7134758a8c61aa961f1",
                    "sha256": "0757c21fc4c955da5351922a234ac300b09a4f87b60684fb27d4e8b2529ce643"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.6.tar.gz",
                "has_sig": false,
                "md5_digest": "246a0f24513aa7134758a8c61aa961f1",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4",
                "size": 44370,
                "upload_time": "2018-11-09T13:47:50",
                "url": "https://files.pythonhosted.org/packages/87/a7/fd8533ba73704b0805ae7050a47a9025355fec95ca0047e93c66fd016442/facenet_sandberg-0.0.6.tar.gz"
            }
        ],
        "0.0.7": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "d848a25d6377466a05988d83a14a25ec",
                    "sha256": "fb5f164a42681553bf4da7546ccd4d91ee15cda2aee2e052c22ecec885c2704d"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.7.tar.gz",
                "has_sig": false,
                "md5_digest": "d848a25d6377466a05988d83a14a25ec",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4",
                "size": 42644,
                "upload_time": "2018-11-09T15:34:50",
                "url": "https://files.pythonhosted.org/packages/e1/23/de9de693401cc3bc5e352a3e63fff4bbdbe1c6b387e856d3b73eb4629253/facenet_sandberg-0.0.7.tar.gz"
            }
        ],
        "0.0.8": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "1584912265b6c3ed9e45eda2a339a16e",
                    "sha256": "57924a03f8ecdf034e105055fd67f33a08172060718aeb3edee063358bc8a35e"
                },
                "downloads": -1,
                "filename": "facenet_sandberg-0.0.8.tar.gz",
                "has_sig": false,
                "md5_digest": "1584912265b6c3ed9e45eda2a339a16e",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.4",
                "size": 42792,
                "upload_time": "2018-11-09T16:30:59",
                "url": "https://files.pythonhosted.org/packages/5a/f4/bbe35ad554c4e3bfd1957d38493d73344c2f25b45b85e9db15b09ef33ad0/facenet_sandberg-0.0.8.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "1584912265b6c3ed9e45eda2a339a16e",
                "sha256": "57924a03f8ecdf034e105055fd67f33a08172060718aeb3edee063358bc8a35e"
            },
            "downloads": -1,
            "filename": "facenet_sandberg-0.0.8.tar.gz",
            "has_sig": false,
            "md5_digest": "1584912265b6c3ed9e45eda2a339a16e",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">=3.4",
            "size": 42792,
            "upload_time": "2018-11-09T16:30:59",
            "url": "https://files.pythonhosted.org/packages/5a/f4/bbe35ad554c4e3bfd1957d38493d73344c2f25b45b85e9db15b09ef33ad0/facenet_sandberg-0.0.8.tar.gz"
        }
    ]
}