人臉檢測+人體檢測C++ Android實作

本博客將實作C++版本的人臉檢測，人臉關鍵點檢測，人體檢測，人臉+人體檢測，推理框架采用TNN，在普通Android手機，CPU和GPU都可以達到實時檢測的效果

人臉檢測+人臉關鍵點檢測+人體檢測Android Demo APP(非原始碼，僅供學習交流)

鏈接: https://pan.baidu.com/s/1By43I1DbMa0gBPLObtPZMQ 提取碼: msnr

尊重原創，轉載請注明出處：https://panjinquan.blog.csdn.net/article/details/120688804

1.專案說明

（1）訓練

訓練代碼請參考：https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB ，一個基于SSD簡化的人臉檢測模型，很輕量化，整個模型僅僅1.7M左右，在普通Android手機都可以實時檢測，

原始代碼使用WiderFace人臉資料集進行訓練，僅支持了人臉檢測，后經鄙人優化后，提高了人臉檢測效果，并支持人臉關鍵點檢測，人體檢測，資料集是WiderFace,VOC和COCO，

（2）端上部署

原始代碼已經支持MNN和NCNN

（3）依賴庫

TNN：https://github.com/Tencent/TNN
OpenCV: Releases - OpenCV （推薦opencv-4.3.0）
OpenCL: Choose & Download Intel? SDK for OpenCL? Applications (GPU的支持)
base-utils：https://github.com/PanJinquan/base-utils (一些檔案和影像處理的相關工具)
拉取子模塊submodule(TNN,base-utils)庫

# pull 3rdparty(TNN,base-utils) submodule
git submodule init
git submodule update

配置OpenCV

推薦opencv-4.3.0

mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local ..
sudo make install

配置OpenCL（可選）

Android系統一般都支持OpenCL，Linux系統可參考如下配置：

# 參考安裝OpenCL： https://blog.csdn.net/qq_28483731/article/details/68235383，作為測驗，安裝`intel cpu版本的OpenCL`即可
# 安裝clinfo，clinfo是一個顯示OpenCL平臺和設備的軟體
sudo apt-get install clinfo
# 安裝依賴
sudo apt install dkms xz-utils openssl libnuma1 libpciaccess0 bc curl libssl-dev lsb-core libicu-dev
sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys 3FA7E0328081BFF6A14DA29AA6A19B38D3D831EF
echo "deb http://download.mono-project.com/repo/debian wheezy main" | sudo tee /etc/apt/sources.list.d/mono-xamarin.list
sudo apt-get update
sudo apt-get install mono-complete
# 在intel官網上下載了intel SDK的tgz檔案，并且解壓
sudo sh install.sh

CMake配置說明

Linux OR Windows測驗，CMakeLists.txt

# TNN set
    set(TNN_OPENCL_ENABLE ON CACHE BOOL "" FORCE)
    set(TNN_CPU_ENABLE ON CACHE BOOL "" FORCE)
    set(TNN_X86_ENABLE ON CACHE BOOL "" FORCE)
    set(TNN_QUANTIZATION_ENABLE OFF CACHE BOOL "" FORCE)
    set(TNN_OPENMP_ENABLE ON CACHE BOOL "" FORCE)  # Multi-Thread
    add_definitions(-DTNN_OPENCL_ENABLE)           # for OpenCL GPU
    add_definitions(-DDEBUG_ON)                    # for WIN/Linux Log
    add_definitions(-DDEBUG_LOG_ON)                # for WIN/Linux Log
    add_definitions(-DDEBUG_IMSHOW_OFF)            # for OpenCV show
    add_definitions(-DPLATFORM_LINUX)
    # add_definitions(-DPLATFORM_WINDOWS)

2. 人臉人體檢測Demo

推理框架使用TNN進行部署，手擼Python轉C++實作人臉人體檢測推理程序，

下面是測驗代碼demo部分代碼

void test_face_person_detector() {
    const int num_thread = 1;
    DeviceType device = CPU;
    // 人臉和關鍵點檢測
    // const char *model_file = (char *) "../data/tnn/face_ldmks/rfb_landm_face_320_320_sim.opt.tnnmodel";
    // const char *proto_file = (char *) "../data/tnn/face_ldmks/rfb_landm_face_320_320_sim.opt.tnnproto";
    // ObjectDetectiobParam model_param = FACE_LANDMARK_MODEL;

    // 人臉+人體檢測
    // const char *model_file = (char *) "../data/tnn/face_person/rfb1.0_face_person_300_300_sim.opt.tnnmodel";
    // const char *proto_file = (char *) "../data/tnn/face_person/rfb1.0_face_person_300_300_sim.opt.tnnproto";
    // ObjectDetectiobParam model_param = FACE_PERSON_MODEL;//模型引數
    // 人臉檢測
    const char *model_file = (char *) "../data/tnn/face/rfb1.0_face_320_320.opt.tnnmodel";
    const char *proto_file = (char *) "../data/tnn/face/rfb1.0_face_320_320.opt.tnnproto";
    ObjectDetectiobParam model_param = FACE_MODEL;//模型引數


    // 設定檢測閾值
    const float scoreThresh = 0.5;
    const float iouThresh = 0.3;
    ObjectDetection *detector = new ObjectDetection(model_file,
                                                    proto_file,
                                                    model_param,
                                                    num_thread,
                                                    device);

    string image_dir = "../data/test_image/person";
    std::vector<string> image_list = get_files_list(image_dir);
    for (string image_path:image_list) {
        cv::Mat bgr_image = cv::imread(image_path);
        if (bgr_image.empty()) continue;
        FrameInfo resultInfo;
        printf("init frame\n");
        // 開始檢測
        detector->detect(bgr_image, &resultInfo, scoreThresh, iouThresh);
        // 可視化代碼
        detector->visualizeResult(bgr_image, &resultInfo);
    }
    delete detector;
    detector = nullptr;
    printf("FINISHED.\n");
}