使用 🍅🍅 Lite.AI.ToolKit C++工具箱来跑SSRNet年龄估计的一些案例, 包含ONNXRuntime C++、MNN和TNN版本。SSRNet的权重文件大小只有 190Kb ，是一个非常轻量级的年龄估计模型。
如果觉得有用，不妨给个Star⭐️🌟支持一下吧~ 🙃🤪🍀
SSRNet C++ 版本的源码包含ONNXRuntime、MNN和TNN三个版本，源码可以在 lite.ai.toolkit 工具箱中找到。本项目主要介绍如何基于 lite.ai.toolkit 工具箱，直接使用SSRNet来跑人脸检测。需要说明的是，本项目是基于MacOS下编译的 liblite.ai.toolkit.v0.1.0.dylib 来实现的，对于使用MacOS的用户，可以直接下载本项目包含的liblite.ai.toolkit.v0.1.0动态库和其他依赖库进行使用。而非MacOS用户，则需要从lite.ai.toolkit 中下载源码进行编译。lite.ai.toolkit c++工具箱目前包含80+流行的开源模型，就不多介绍了，只是平时顺手捏的，整合了自己学习过程中接触到的一些模型，感兴趣的同学可以去看看。
ONNXRuntime C++、MNN和TNN版本的推理实现均已测试通过，欢迎白嫖~
可以从我提供的链接下载 (Baidu Drive code: 8gin) , 也可以从本直接仓库下载。
| Class | Pretrained ONNX Files | Rename or Converted From (Repo) | Size |
|---|---|---|---|
| lite::cv::face::attr::SSRNet | ssrnet.onnx | SSR_Net... | 190Kb |
MNN模型文件下载地址，(Baidu Drive code: 9v63), 也可以从本直接仓库下载。
| Class | Pretrained MNN Files | Rename or Converted From (Repo) | Size |
|---|---|---|---|
| lite::mnn::cv::face::attr::SSRNet | ssrnet.mnn | SSR_Net... | 190Kb |
TNN模型文件下载地址，(Baidu Drive code: 6o6k), 也可以从本直接仓库下载。
| Class | Pretrained TNN Files | Rename or Converted From (Repo) | Size |
|---|---|---|---|
| lite::tnn::cv::face::attr::SSRNet | ssrnet.opt.tnnproto&tnnmodel | SSR_Net... | 190Kb |
在lite.ai.toolkit 中，SSRNet的实现类为：
class LITE_EXPORTS lite::cv::face::attr::SSRNet;
class LITE_EXPORTS lite::mnn::cv::face::attr::SSRNet;
class LITE_EXPORTS lite::tnn::cv::face::attr::SSRNet;该类型目前包含1公共接口detect用于进行年龄检测。
public:
void detect(const cv::Mat &mat, types::Age &age);detect接口的输入参数说明：
- mat: cv::Mat类型，BGR格式，一张包含人脸头部的图片（不包含过多的背景）。
- age: types::Age，包含被检测到的年龄;
#include "lite/lite.h"
static void test_default()
{
std::string onnx_path = "../hub/onnx/cv/ssrnet.onnx";
std::string test_img_path = "../resources/1.png";
std::string save_img_path = "../logs/1.jpg";
auto *ssrnet = new lite::cv::face::attr::SSRNet(onnx_path);
lite::types::Age age;
cv::Mat img_bgr = cv::imread(test_img_path);
ssrnet->detect(img_bgr, age);
lite::utils::draw_age_inplace(img_bgr, age);
cv::imwrite(save_img_path, img_bgr);
std::cout << "Default Version Done! Detected SSRNet Age: " << age.age << std::endl;
delete ssrnet;
}#include "lite/lite.h"
static void test_mnn()
{
#ifdef ENABLE_MNN
std::string mnn_path = "../hub/mnn/cv/ssrnet.mnn";
std::string test_img_path = "../resources/3.png";
std::string save_img_path = "../logs/3_mnn.jpg";
auto *ssrnet = new lite::mnn::cv::face::attr::SSRNet(mnn_path);
lite::types::Age age;
cv::Mat img_bgr = cv::imread(test_img_path);
ssrnet->detect(img_bgr, age);
lite::utils::draw_age_inplace(img_bgr, age);
cv::imwrite(save_img_path, img_bgr);
std::cout << "MNN Version Done! Detected SSRNet Age: " << age.age << std::endl;
delete ssrnet;
#endif
}#include "lite/lite.h"
static void test_tnn()
{
#ifdef ENABLE_TNN
std::string proto_path = "../hub/tnn/cv/ssrnet.opt.tnnproto";
std::string model_path = "../hub/tnn/cv/ssrnet.opt.tnnmodel";
std::string test_img_path = "../resources/4.png";
std::string save_img_path = "../logs/4_tnn.jpg";
auto *ssrnet = new lite::tnn::cv::face::attr::SSRNet(proto_path, model_path);
lite::types::Age age;
cv::Mat img_bgr = cv::imread(test_img_path);
ssrnet->detect(img_bgr, age);
lite::utils::draw_age_inplace(img_bgr, age);
cv::imwrite(save_img_path, img_bgr);
std::cout << "TNN Version Done! Detected SSRNet Age: " << age.age << std::endl;
delete ssrnet;
#endif
}- 输出结果为:
在MacOS下可以直接编译运行本项目，无需下载其他依赖库。其他系统则需要从lite.ai.toolkit 中下载源码先编译lite.ai.toolkit.v0.1.0动态库。
git clone --depth=1 https://github.com/DefTruth/ssrnet.lite.ai.toolkit.git
cd ssrnet.lite.ai.toolkit
sh ./build.sh- CMakeLists.txt设置
cmake_minimum_required(VERSION 3.17)
project(ssrnet.lite.ai.toolkit)
set(CMAKE_CXX_STANDARD 11)
# setting up lite.ai.toolkit
set(LITE_AI_DIR ${CMAKE_SOURCE_DIR}/lite.ai.toolkit)
set(LITE_AI_INCLUDE_DIR ${LITE_AI_DIR}/include)
set(LITE_AI_LIBRARY_DIR ${LITE_AI_DIR}/lib)
include_directories(${LITE_AI_INCLUDE_DIR})
link_directories(${LITE_AI_LIBRARY_DIR})
set(OpenCV_LIBS
opencv_highgui
opencv_core
opencv_imgcodecs
opencv_imgproc
opencv_video
opencv_videoio
)
# add your executable
set(EXECUTABLE_OUTPUT_PATH ${CMAKE_SOURCE_DIR}/examples/build)
add_executable(lite_ssrnet examples/test_lite_ssrnet.cpp)
target_link_libraries(lite_ssrnet
lite.ai.toolkit
onnxruntime
MNN # need, if built lite.ai.toolkit with ENABLE_MNN=ON, default OFF
ncnn # need, if built lite.ai.toolkit with ENABLE_NCNN=ON, default OFF
TNN # need, if built lite.ai.toolkit with ENABLE_TNN=ON, default OFF
${OpenCV_LIBS}) # link lite.ai.toolkit & other libs.- building && testing information:
[ 50%] Building CXX object CMakeFiles/lite_ssrnet.dir/examples/test_lite_ssrnet.cpp.o
[100%] Linking CXX executable lite_ssrnet
[100%] Built target lite_ssrnet
Testing Start ...
LITEORT_DEBUG LogId: ../hub/onnx/cv/ssrnet.onnx
=============== Input-Dims ==============
input_node_dims: 1
input_node_dims: 3
input_node_dims: 64
input_node_dims: 64
=============== Output-Dims ==============
Output: 0 Name: age Dim: 0 :1
========================================
Default Version Done! Detected SSRNet Age: 35.567
LITEORT_DEBUG LogId: ../hub/onnx/cv/ssrnet.onnx
=============== Input-Dims ==============
input_node_dims: 1
input_node_dims: 3
input_node_dims: 64
input_node_dims: 64
=============== Output-Dims ==============
Output: 0 Name: age Dim: 0 :1
========================================
ONNXRuntime Version Done! Detected SSRNet Age: 27.4245
LITEMNN_DEBUG LogId: ../hub/mnn/cv/ssrnet.mnn
=============== Input-Dims ==============
**Tensor shape**: 1, 3, 64, 64,
Dimension Type: (CAFFE/PyTorch/ONNX)NCHW
=============== Output-Dims ==============
getSessionOutputAll done!
Output: age: **Tensor shape**: 1,
========================================
MNN Version Done! Detected SSRNet Age: 33.37
LITETNN_DEBUG LogId: ../hub/tnn/cv/ssrnet.opt.tnnproto
=============== Input-Dims ==============
input: [1 3 64 64 ]
Input Data Format: NCHW
=============== Output-Dims ==============
age: [1 ]
========================================
TNN Version Done! Detected SSRNet Age: 23.7442
Testing Successful !