vitis::ai::RetinaFace

Base class for detecting the position,score and landmark of faces in the input image (cv::Mat).

Input is an image (cv::Mat).

Output is a vector of position and score for faces in the input image.

Sample code:

auto image = cv::imread("sample_retinaface.jpg");
auto network = vitis::ai::RetinaFace::create(
               "retinaface",
               true);
auto result = network->run(image);
for (auto i = 0u; i < result.bboxes.size(); ++i) {
   auto score = result.bboxes[i].score;
   auto x = result.bboxes[i].x * image.cols;
   auto y = result.bboxes[i].y * image.rows;
   auto width = result.bboxes[i].width * image.cols;
   auto height = result.bboxes[i].height * image.rows;
   auto landmark = results.landmarks[i];
   for (auto j = 0; j < 5; ++j) {
     auto px = landmark[j].first * image.cols;
     auto py = landmark[j].second * image.rows;
   }
}

Display of the model results:

Image sample_retinaface_result.jpg — Figure 1: result image

Quick Function Reference

The following table lists all the functions defined in the vitis::ai::RetinaFace class:

Table 1. Quick Function Reference
Type	Name	Arguments
std::unique_ptr< `RetinaFace` >	create	const std::string & model_name bool need_preprocess
std::unique_ptr< `RetinaFace` >	create	void
`RetinaFaceResult`	run	const cv::Mat & img
std::vector< `RetinaFaceResult` >	run	const std::vector< cv::Mat > & imgs
std::vector< `RetinaFaceResult` >	run	const std::vector< vart::xrt_bo_t > & input_bos

create

Factory function to get an instance of derived classes of class RetinaFace.

Prototype

std::unique_ptr< RetinaFace > create(const std::string &model_name, bool need_preprocess=true);

Parameters

The following table lists the create function arguments.

Table 2. create Arguments
Type	Name	Description
const std::string &	model_name	Model name
bool	need_preprocess	Normalize with mean/scale or not, default value is true.

Returns

An instance of RetinaFace class.

create

Prototype

std::unique_ptr< RetinaFace > create(const std::string &model_name, xir::Attrs *attrs, bool need_preprocess=true);

run

Function to get running result of the retinaface network.

Prototype


            RetinaFaceResult run(const cv::Mat &img)=0;

Parameters

The following table lists the run function arguments.

Table 3. run Arguments
Type	Name	Description
const cv::Mat &	img	Input Data ,input image (cv::Mat) need to be resized to InputWidth and InputHeight required by the network.

Returns

The detection result of the face detect network , filtered by score >= det_threshold

run

Function to get running results of the retinaface neuron network in batch mode.

Prototype

std::vector< RetinaFaceResult > run(const std::vector< cv::Mat > &imgs)=0;

Parameters

The following table lists the run function arguments.

Table 4. run Arguments
Type	Name	Description
const std::vector< cv::Mat > &	imgs	Input data of input images (std:vector<cv::Mat>). The size of input images equals batch size obtained by get_input_batch. The input images need to be resized to InputWidth and InputHeight required by the network.

Returns

The vector of RetinaFaceResult.

run

Function to get running results of the retina neuron network in batch mode , used to receive user's xrt_bo to support zero copy.

Prototype

std::vector< RetinaFaceResult > run(const std::vector< vart::xrt_bo_t > &input_bos)=0;

Parameters

The following table lists the run function arguments.

Table 5. run Arguments
Type	Name	Description
const std::vector< vart::xrt_bo_t > &	input_bos	The vector of vart::xrt_bo_t.

Returns

The vector of RetinaFacesResult.