Files
MAXINE-AR-SDK/samples/GazeRedirect/GazeEngine.h
Joy D'Souza ca10ac3b39 v0.8.2.0 Release
v0.8.2.0 Release
2023-01-06 10:03:38 -08:00

209 lines
6.8 KiB
C++

/*###############################################################################
#
# Copyright 2020 NVIDIA Corporation
#
# Permission is hereby granted, free of charge, to any person obtaining a copy of
# this software and associated documentation files (the "Software"), to deal in
# the Software without restriction, including without limitation the rights to
# use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
# the Software, and to permit persons to whom the Software is furnished to do so,
# subject to the following conditions:
#
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
#
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
# FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
# COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
# IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
# CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
#
###############################################################################*/
#ifndef __GAZE_ENGINE__
#define __GAZE_ENGINE__
#include <random>
#include "FeatureVertexName.h"
#include "nvAR.h"
#include "nvCVOpenCV.h"
class KalmanFilter1D {
private:
float Q_; // Covariance of the process noise
float xhat_; // Current prediction
float xhatminus_; // Previous prediction
float P_; // Estimated accuracy of xhat_
float Pminus_; // Previous P_
float K_; // Kalman gain
float R_; // Covariance of the observation noise
bool bFirstUse;
public:
KalmanFilter1D() { reset(); }
KalmanFilter1D(float Q, float R) { reset(Q, R); }
void reset() {
R_ = 0.005f * 0.005f;
Q_ = 1e-5f;
xhat_ = 0.0f;
xhatminus_ = 0.0f;
P_ = 1;
bFirstUse = true;
Pminus_ = 0.0f;
K_ = 0.0f;
}
void reset(float Q, float R) {
reset();
Q_ = Q;
R_ = R;
}
float update(float val) {
if (bFirstUse) {
xhat_ = val;
bFirstUse = false;
}
xhatminus_ = xhat_;
Pminus_ = P_ + Q_;
K_ = Pminus_ / (Pminus_ + R_);
xhat_ = xhatminus_ + K_ * (val - xhatminus_);
P_ = (1 - K_) * Pminus_;
return xhat_;
}
};
bool CheckResult(NvCV_Status nvErr, unsigned line);
#define BAIL_IF_ERR(err) \
do { \
if (0 != (err)) { \
goto bail; \
} \
} while (0)
#define BAIL_IF_NVERR(nvErr, err, code) \
do { \
if (!CheckResult(nvErr, __LINE__)) { \
err = code; \
goto bail; \
} \
} while (0)
typedef struct LandmarksProperties {
int numPoints;
float confidence_threshold;
} LandmarksProperties;
/********************************************************************************
* GazeEngine
********************************************************************************/
class GazeEngine {
public:
enum Err { errNone, errGeneral, errRun, errInitialization, errRead, errEffect, errParameter };
int input_image_width, input_image_height, input_image_pitch;
const LandmarksProperties LANDMARKS_INFO[2] = {{68, 15.0f}, // number of landmark points, confidence threshold value
{126, 5.0f}};
void setInputImageWidth(int width) { input_image_width = width; }
void setInputImageHeight(int height) { input_image_height = height; }
Err createGazeRedirectionFeature(const char* modelPath, unsigned int _batchSize = 1);
void destroyGazeRedirectionFeature();
Err initGazeRedirectionIOParams();
unsigned findFaceBoxes();
NvAR_Rect* getLargestBox();
NvCV_Status findLandmarks();
NvAR_BBoxes* getBoundingBoxes();
/**
* Landmarks corresponding to facial keypoints
*
* @returns Pointer to the landmarks array
*/
NvAR_Point2f* getLandmarks();
/**
* Output landmarks corresponding to the redirected eyes from the gaze redirection network
*
* @returns Pointer to the landmarks array
*/
NvAR_Point2f* getGazeOutputLandmarks();
NvAR_Quaternion* getPose();
float* getHeadTranslation();
float* getGazeVector();
float* getLandmarksConfidence();
float getAverageLandmarksConfidence();
void enlargeAndSquarifyImageBox(float enlarge, NvAR_Rect& box, int FLAG_variant);
unsigned findLargestFaceBox(NvAR_Rect& faceBox, int variant = 0);
unsigned acquireFaceBox(cv::Mat& src, NvAR_Rect& faceBox, int variant = 0);
unsigned acquireFaceBoxAndLandmarks(cv::Mat& src, NvAR_Point2f* refMarks, NvAR_Rect& faceBox);
Err acquireGazeRedirection(cv::Mat& frame, cv::Mat& outputFrame);
NvAR_RenderingParams* getRenderingParams();
void setFaceStabilization(bool);
Err setNumLandmarks(int);
void setGazeRedirect(bool _bGazeRedirect);
void setUseCudaGraph(bool _bUseCudaGraph);
void setEyeSizeSensitivity(unsigned);
int getNumLandmarks() { return numLandmarks; }
int getNumGazeOutputLandmarks() { return num_output_landmarks; }
void DrawPose(const cv::Mat& src, const NvAR_Quaternion* pose) const;
std::array<float, 2> GetAverageLandmarkPositionInGlSpace() const;
void DrawEstimatedGaze(const cv::Mat& src);
NvAR_Point3f* getGazeDirectionPoints();
NvCVImage inputImageBuffer{}, tmpImage{}, outputImageBuffer{};
NvAR_FeatureHandle faceDetectHandle{}, landmarkDetectHandle{}, gazeRedirectHandle{};
std::vector<NvAR_Point2f> facial_landmarks;
std::vector<NvAR_Point2f> gaze_output_landmarks;
std::vector<float> facial_landmarks_confidence;
NvAR_Point3f gaze_direction[2] = {{0.f, 0.f, 0.f}};
NvAR_Quaternion head_pose;
float gaze_angles_vector[2] = {0.f};
float head_translation[3] = {0.f};
NvAR_RenderingParams* rendering_params{};
CUstream stream{};
std::vector<NvAR_Rect> output_bbox_data;
std::vector<float> output_bbox_conf_data;
NvAR_BBoxes output_bboxes{};
int batchSize;
std::mt19937 ran;
int numLandmarks;
int num_output_landmarks;
int eyeSizeSensitivity;
float confidenceThreshold;
std::string face_model;
bool bStabilizeFace;
bool bUseOTAU;
bool bGazeRedirect;
bool bUseCudaGraph;
char *fdOTAModelPath, *ldOTAModelPath;
GazeEngine() {
batchSize = 1;
bStabilizeFace = true;
bGazeRedirect = true;
bUseCudaGraph = true;
numLandmarks = LANDMARKS_INFO[0].numPoints;
num_output_landmarks = 12;
confidenceThreshold = LANDMARKS_INFO[0].confidence_threshold;
input_image_width = 640;
input_image_height = 480;
input_image_pitch = 3 * input_image_width * sizeof(unsigned char); // RGB
bUseOTAU = false;
fdOTAModelPath = NULL;
ldOTAModelPath = NULL;
eyeSizeSensitivity = 3;
}
};
#endif