Python实现多人脸识别系统

冷不防 2023-10-15 14:38 22阅读 0赞

## [资源下载地址][Link 1]：https://download.csdn.net/download/sheziqiong/88284343   [资源下载地址][Link 1]：https://download.csdn.net/download/sheziqiong/88284343   Face recognition from camera with Dlib ##

### Introduction ###

调用摄像头进行人脸识别, 支持多张人脸同时识别

1.  Tkinter 人脸录入界面, 支持录入时设置 (中文) 姓名
    
    ![652b3c600456fe4d89ae7a48aee6e91e.png][]\{.align-center  
    width=“1000px”\}
2.  简单的 OpenCV 摄像头人脸录入界面
    
    ![a0cfd94099a8ab0722f8c9d1e5f9601c.png][]\{.align-center  
    width=“1000px”\}
    
    离摄像头过近, 人脸超出摄像头范围时, 会有 “OUT OF RANGE” 提醒
    
    ![外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传][20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F33c1df21d1d9f8ef107e605042212dad.writebug_pos_id_img-CFka6398-1693554083910]\{.align-center  
    width=“1000px”\}
3.  提取特征建立人脸数据库
4.  利用摄像头进行人脸识别
    
    face\_reco\_from\_camera.py, 对于每一帧都做检测识别
    
    ![外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传][20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F0b06e55d27372b3f47ac143998d4ea20.writebug_pos_id_img-twoXknAm-1693554083911]\{.align-center width=“1000px”\}
    
    face\_reco\_from\_camera\_single\_face.py, 对于人脸<=1,  
    只有新人脸出现才进行再识别来提高 FPS
    
    ![d73f40da344961c8641d13fa6fe6509d.png][]\{.align-center  
    width=“1000px”\}
    
    face\_reco\_from\_camera\_ot.py, 利用 OT 来实现再识别提高 FPS
    
    ![外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传][20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F3cc3c9317a6325b9f756ae9ae308059c.writebug_pos_id_img-l1wFhpFB-1693554083912]\{.align-center  
    width=“1000px”\}
    
    定制显示名字, 可以写中文
    
    ![外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传][20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F6af0b809f698fe93139394744d66489d.writebug_pos_id_img-RNpPYUhW-1693554083912]\{.align-center  
    width=“1000px”\}

**关于精度** :

*  当使用 0.6 的距离阈值时，dlib 模型在标准 LFW 人脸识别基准上获得 99.38% 的精度。

**关于算法**

*  基于 Residual Neural Network / 残差网络的 CNN 模型;
 *  This model is a ResNet network with 29 conv layers.

本质上是 ResNet-34 网络的一个版本，来自 He，Zhang，Ren 和 Sun 的论文 Deep Residual Learning for Image Recognition，去掉了几层，每层过滤器的数量减少了一半。

### 概述 ###

此项目中人脸识别的实现流程 (no OT, 每一帧都进行检测 + 识别)

![1e19eb240228fce700a157136719515e.png][]\{.align-center width=“1000px”\}

实现流程 (with OT, 初始帧进行检测 + 识别, 后续帧检测 + 质心跟踪) / OT used:

![74034cf675ef25a9ca2812dd8afe24cb.png][]\{.align-center  
width=“1000px”\}

如果利用 OT 来跟踪, 可以大大提高 FPS,因为做识别时候需要提取特征描述子的耗时很多

### Steps ###

1.  / Git clone source code

```{.bash}
    git clone https://git.writebug.com/goodwill/DlibFaceRecognition.git

2. 安装依赖库
    
       ```{.bash}
       pip install -r requirements.txt

1.  进行人脸信息采集录入,
    
        # Install Tkinter
        sudo apt-get install python3-tk python3-pil python3-pil.imagetk
        
        python3 get_faces_from_camera_tkinter.py
2.  进行人脸信息采集录入, OpenCV GUI
    
        python3 get_face_from_camera.py
3.  提取所有录入人脸数据存入 `features_all.csv`
    
        python3 features_extraction_to_csv.py
4.  调用摄像头进行实时人脸识别
    
        python3 face_reco_from_camera.py
5.  对于人脸数<=1, 调用摄像头进行实时人脸识别
    
        python3 face_reco_from_camera_single_face.py
6.  利用 OT 算法, 调用摄像头进行实时人脸识别
    
        python3 face_reco_from_camera_ot.py

### 源代码相关 ###

代码结构 :

.
    ├── get_faces_from_camera.py                # Step 1. Face register GUI with OpenCV
    ├── get_faces_from_camera_tkinter.py                # Step 1. Face register GUI with Tkinter
    ├── features_extraction_to_csv.py           # Step 2. Feature extraction
    ├── face_reco_from_camera.py                # Step 3. Face recognizer
    ├── face_reco_from_camera_single_face.py            # Step 3. Face recognizer for single person
    ├── face_reco_from_camera_ot.py                     # Step 3. Face recognizer with OT
    ├── face_descriptor_from_camera.py          # Face descriptor computation
    ├── how_to_use_camera.py                    # Use the default camera by opencv
    ├── data
    │ ├── data_dlib                           # Dlib's model
    │ │ ├── dlib_face_recognition_resnet_model_v1.dat
    │ │ └── shape_predictor_68_face_landmarks.dat
    │ ├── data_faces_from_camera                      # Face images captured from camera (will generate after step 1)
    │ │ ├── person_1
    │ │ │ ├── img_face_1.jpg
    │ │ │ └── img_face_2.jpg
    │ │ └── person_2
    │ │     └── img_face_1.jpg
    │ │     └── img_face_2.jpg
    │ └── features_all.csv                        # CSV to save all the features of known faces (will generate after step 2)
    ├── README.rst
    └── requirements.txt                        # Some python packages needed

用到的 Dlib 相关模型函数

1.  Dlib 正向人脸检测器  
    `<class 'dlib.dlib.rectangles'>` / Dlib frontal face detector
    
        detector = dlib.get_frontal_face_detector()
        faces = detector(img_gray, 0)
2.  Dlib 人脸 landmark 特征点检测器, output:  
    `<class 'dlib.dlib.full_object_detection'>` / Dlib face landmark  
    predictor, will use `shape_predictor_68_face_landmarks.dat`
    
        # This is trained on the ibug 300-W dataset (https://ibug.doc.ic.ac.uk/resources/facial-point-annotations/)
        # Also note that this model file is designed for use with dlib's HOG face detector.
        # That is, it expects the bounding boxes from the face detector to be aligned a certain way,
        the way dlib's HOG face detector does it.
        # It won't work as well when used with a face detector that produces differently aligned boxes,
        # such as the CNN based mmod_human_face_detector.dat face detector.
        
        predictor = dlib.shape_predictor("data/data_dlib/shape_predictor_68_face_landmarks.dat")
        shape = predictor(img_rd, faces[i])
3.  Dlib 特征描述子 / Face recognition model, the object maps human  
    faces into 128D vectors
    
        face_rec = dlib.face_recognition_model_v1("data/data_dlib/dlib_face_recognition_resnet_model_v1.dat")

Python 源码介绍如下

*  `get_face_from_camera.py`:
    
    人脸信息采集录入
    
     *  请注意存储人脸图片时, 矩形框不要超出摄像头范围,  
        要不然无法保存到本地;
     *  超出会有 “out of range” 的提醒;
 *  `get_faces_from_camera_tkinter.py`:
    
    进行人脸信息采集录入 Tkinter GUI
 *  `features_extraction_to_csv.py`:
    
    从上一步存下来的图像文件中, 提取人脸数据存入 CSV  
    from face images saved in step 1;
    
     *  会生成一个存储所有特征人脸数据的 `features_all.csv`
     *  Size: `n*129` , n means n faces you registered and 129 means  
        face name + 128D features of this face
 *  `face_reco_from_camera.py`:
    
    这一步将调用摄像头进行实时人脸识别;  
    real-time face recognition;
    
     *  将捕获到的人脸数据和之前存的人脸数据进行对比计算欧式距离,  
        由此判断是否是同一个人;
     *  Compare the faces captured from camera with the faces you have  
        registered which are saved in `features_all.csv`;
 *  `face_reco_from_camera_single_face.py`:
    
    针对于人脸数 <=1 的场景, 区别于 `face_reco_from_camera.py`  
    (对每一帧都进行检测 + 识别), 只有人脸出现的时候进行识别;
 *  `face_reco_from_camera_ot.py`:
    
    只会对初始帧做检测 + 识别, 对后续帧做检测 + 质心跟踪;
 *  (optional) `face_descriptor_from_camera.py`
    
    调用摄像头进行实时特征描述子计算;  
    computation;

[资源下载地址][Link 1]：https://download.csdn.net/download/sheziqiong/88284343  
[资源下载地址][Link 1]：https://download.csdn.net/download/sheziqiong/88284343

[Link 1]: https://download.csdn.net/download/sheziqiong/88284343
[652b3c600456fe4d89ae7a48aee6e91e.png]: https://img-blog.csdnimg.cn/img_convert/652b3c600456fe4d89ae7a48aee6e91e.png
[a0cfd94099a8ab0722f8c9d1e5f9601c.png]: https://img-blog.csdnimg.cn/img_convert/a0cfd94099a8ab0722f8c9d1e5f9601c.png
[20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F33c1df21d1d9f8ef107e605042212dad.writebug_pos_id_img-CFka6398-1693554083910]: https://img-home.csdnimg.cn/images/20230724024159.png?origin_url=https%3A%2F%2Fwww.writebug.com%2Fmyres%2Fstatic%2Fuploads%2F2021%2F12%2F26%2F33c1df21d1d9f8ef107e605042212dad.writebug&pos_id=img-CFka6398-1693554083910
[20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F0b06e55d27372b3f47ac143998d4ea20.writebug_pos_id_img-twoXknAm-1693554083911]: https://img-home.csdnimg.cn/images/20230724024159.png?origin_url=https%3A%2F%2Fwww.writebug.com%2Fmyres%2Fstatic%2Fuploads%2F2021%2F12%2F26%2F0b06e55d27372b3f47ac143998d4ea20.writebug&pos_id=img-twoXknAm-1693554083911
[d73f40da344961c8641d13fa6fe6509d.png]: https://img-blog.csdnimg.cn/img_convert/d73f40da344961c8641d13fa6fe6509d.png
[20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F3cc3c9317a6325b9f756ae9ae308059c.writebug_pos_id_img-l1wFhpFB-1693554083912]: https://img-home.csdnimg.cn/images/20230724024159.png?origin_url=https%3A%2F%2Fwww.writebug.com%2Fmyres%2Fstatic%2Fuploads%2F2021%2F12%2F26%2F3cc3c9317a6325b9f756ae9ae308059c.writebug&pos_id=img-l1wFhpFB-1693554083912
[20230724024159.png_origin_url_https_3A_2F_2Fwww.writebug.com_2Fmyres_2Fstatic_2Fuploads_2F2021_2F12_2F26_2F6af0b809f698fe93139394744d66489d.writebug_pos_id_img-RNpPYUhW-1693554083912]: https://img-home.csdnimg.cn/images/20230724024159.png?origin_url=https%3A%2F%2Fwww.writebug.com%2Fmyres%2Fstatic%2Fuploads%2F2021%2F12%2F26%2F6af0b809f698fe93139394744d66489d.writebug&pos_id=img-RNpPYUhW-1693554083912
[1e19eb240228fce700a157136719515e.png]: https://img-blog.csdnimg.cn/img_convert/1e19eb240228fce700a157136719515e.png
[74034cf675ef25a9ca2812dd8afe24cb.png]: https://img-blog.csdnimg.cn/img_convert/74034cf675ef25a9ca2812dd8afe24cb.png