snipplets.dev/projects/OpenVINO/C++/infer.cc

#include "infer.hpp"

Inf::Inf(const std::string &model_path, const float &model_probability, const float &model_NMS) {
  input_shape = cv::Size(640, 640);
  probability = model_probability;
  NMS = model_NMS;
  init(model_path);
};

Inf::Inf(const std::string &model_path, const cv::Size model_input_shape, const float &model_probability, const float &model_NMS) {
  input_shape = model_input_shape;
  probability = model_probability;
  NMS = model_NMS;
  init(model_path);
};

void Inf::init(const std::string &model_path) {
  ov::Core core;
  std::shared_ptr<ov::Model> model = core.read_model(model_path);

  // Если модель имеет динамические формы,
  // изменяем модель в соответствиии с указанной формой
  if (model->is_dynamic()) {
    model->reshape({1, 3, static_cast<long int>(input_shape.height), static_cast<long int>(input_shape.width)});
  }

  // Настройка предварительной обработки для модели
  ov::preprocess::PrePostProcessor ppp = ov::preprocess::PrePostProcessor(model);
  ppp.input().tensor().set_element_type(ov::element::u8).set_layout("NHWC").set_color_format(ov::preprocess::ColorFormat::BGR);
  ppp.input()
      .preprocess()
      .convert_element_type(ov::element::f32)
      .convert_color(ov::preprocess::ColorFormat::RGB)
      .scale({255, 255, 255});
  ppp.input().model().set_layout("NCHW");
  ppp.output().tensor().set_element_type(ov::element::f32);
  model = ppp.build();

  compiled_model = core.compile_model(model, "AUTO");
  inference_request = compiled_model.create_infer_request();

  const std::vector<ov::Output<ov::Node>> inputs = model->inputs();
  const ov::Shape in_shape = inputs[0].get_shape();
  input_shape = cv::Size2f(in_shape[2], in_shape[1]);

  const std::vector<ov::Output<ov::Node>> outputs = model->outputs();
  const ov::Shape out_shape = outputs[0].get_shape();
  output_shape = cv::Size(out_shape[2], out_shape[1]);
};

void Inf::pre(const cv::Mat &frame) {
  cv::Mat resized_frame;
  cv::resize(frame, resized_frame, input_shape, 0, 0, cv::INTER_AREA); // Resize the frame to match the model input shape

  // Calculate scaling factor
  scale_factor.x = static_cast<float>(frame.cols / input_shape.width);
  scale_factor.y = static_cast<float>(frame.rows / input_shape.height);

  float *input_data = (float *)resized_frame.data; // Get pointer to resized frame data
  const ov::Tensor input_tensor =
      ov::Tensor(compiled_model.input().get_element_type(), compiled_model.input().get_shape(), input_data);
  inference_request.set_input_tensor(input_tensor); // Set input tensor for inference
};

void Inf::post(cv::Mat &frame) {
  std::vector<int> class_list;
  std::vector<float> confidence_list;
  std::vector<cv::Rect> box_list;

  const float *detections = inference_request.get_output_tensor().data<const float>();
  const cv::Mat detection_outputs(output_shape, CV_32F, (float *)detections);

  for (int i = 0; i < detection_outputs.cols; ++i) {
    const cv::Mat classes_scores = detection_outputs.col(i).rowRange(4, detection_outputs.rows);

    cv::Point class_id;
    double score;
    cv::minMaxLoc(classes_scores, nullptr, &score, nullptr, &class_id);

    if (score > probability) {
      class_list.push_back(class_id.y);
      confidence_list.push_back(score);

      const float x = detection_outputs.at<float>(0, i);
      const float y = detection_outputs.at<float>(1, i);
      const float w = detection_outputs.at<float>(2, i);
      const float h = detection_outputs.at<float>(3, i);

      cv::Rect box;
      box.x = static_cast<int>(x);
      box.y = static_cast<int>(y);
      box.width = static_cast<int>(w);
      box.height = static_cast<int>(h);
      box_list.push_back(box);
    }
  }

  std::vector<int> NMS_result;
  cv::dnn::NMSBoxes(box_list, confidence_list, probability, NMS, NMS_result);

  for (int i = 0; i < NMS_result.size(); ++i) {
    Detection result;
    const unsigned short id = NMS_result[i];

    result.class_id = class_list[id];
    result.probability = confidence_list[id];
    result.box = GetBoundingBox(box_list[id]);

    DrawDetectedObject(frame, result);
  }
};

void Inf::inference(cv::Mat &frame) {
  pre(frame);
  inference_request.infer();
  post(frame);
};

cv::Rect Inf::GetBoundingBox(const cv::Rect &src) const {
  cv::Rect box = src;
  box.x = (box.x - box.width / 2) * scale_factor.x;
  box.y = (box.y - box.height / 2) * scale_factor.y;
  box.width *= scale_factor.x;
  box.height *= scale_factor.y;
  return box;
}

void Inf::DrawDetectedObject(cv::Mat &frame, const Detection &detection) const {
    const cv::Rect &box = detection.box;
    const float &confidence = detection.probability;
    const int &class_id = detection.class_id;

    const cv::Scalar &color = cv::Scalar(0, 0, 180);

    cv::rectangle(frame, cv::Point(box.x, box.y), cv::Point(box.x + box.width, box.y + box.height), color, 3);
}