C++代码
#include <pybind11/pybind11.h>
#include<opencv2/opencv.hpp>
#include<pybind11/numpy.h>
using namespace std;
using namespace cv;
namespace py = pybind11;
void BilinearInsert(Mat& src, Mat& dst, float ux, float uy, int i, int j)
{
auto Abs = [&](float f) {
return f > 0 ? f : -f;
};
int height = src.rows-1;
int width = src.cols-1;
int c = src.channels();
if (c == 3)
{
//存储图像得浮点坐标
CvPoint2D32f uv;
CvPoint3D32f f1;
CvPoint3D32f f2;
//取整数
int iu = (int)ux;
int iv = (int)uy;
if (iu > width)
{
iu = width;
}
if (iv > height)
{
iv = height;
}
uv.x = iu + 1;
uv.y = iv + 1;
//step图象像素行的实际宽度 三个通道进行计算(0 , 1 2 三通道)
/*cout << "src.data" << src.data << std::endl;
//C++ unsigned char ＊是表示无符号字符指针的意思
//unsigned若省略后一个关键字，大多数编译器都会认为是unsigned int。
/*cout << "src.data.type" << typeid(src.data).name() << std::endl;*/
f1.x = ((uchar*)(src.data + src.step * iv))[iu * 3] * (1 - Abs(uv.x - iu)) + \
((uchar*)(src.data + src.step * iv))[(iu + 1) * 3] * (uv.x - iu);
f1.y = ((uchar*)(src.data + src.step * iv))[iu * 3 + 1] * (1 - Abs(uv.x - iu)) + \
((uchar*)(src.data + src.step * iv))[(iu + 1) * 3 + 1] * (uv.x - iu);
f1.z = ((uchar*)(src.data + src.step * iv))[iu * 3 + 2] * (1 - Abs(uv.x - iu)) + \
((uchar*)(src.data + src.step * iv))[(iu + 1) * 3 + 2] * (uv.x - iu);
f2.x = ((uchar*)(src.data + src.step * (iv + 1)))[iu * 3] * (1 - Abs(uv.x - iu)) + \
/*cout << "f1.x" << f1.x << std::endl;*/
((uchar*)(src.data + src.step * (iv + 1)))[(iu + 1) * 3] * (uv.x - iu);
f2.y = ((uchar*)(src.data + src.step * (iv + 1)))[iu * 3 + 1] * (1 - Abs(uv.x - iu)) + \
((uchar*)(src.data + src.step * (iv + 1)))[(iu + 1) * 3 + 1] * (uv.x - iu);
f2.z = ((uchar*)(src.data + src.step * (iv + 1)))[iu * 3 + 2] * (1 - Abs(uv.x - iu)) + \
((uchar*)(src.data + src.step * (iv + 1)))[(iu + 1) * 3 + 2] * (uv.x - iu);
//step指的字节数,一个图像是3通道，每个通道是16位，那么
//其elemsize就是3* （16 / 8） = 6，即每个元素占6个字节
((uchar*)(dst.data + dst.step * j))[i * 3] = f1.x * (1 - Abs(uv.y - iv)) + f2.x * (Abs(uv.y - iv)); //三个通道进行赋值
((uchar*)(dst.data + dst.step * j))[i * 3 + 1] = f1.y * (1 - Abs(uv.y - iv)) + f2.y * (Abs(uv.y - iv));
((uchar*)(dst.data + dst.step * j))[i * 3 + 2] = f1.z * (1 - Abs(uv.y - iv)) + f2.z * (Abs(uv.y - iv));
}
}
Mat LocalTranslationWarp_Face(Mat& img, int warpX, int warpY, int endX, int endY, float radius)
{
Mat dst = img.clone();
//平移距离
float ddradius = radius * radius;
//计算|m-c|^2
size_t mc = (endX - warpX) * (endX - warpX) + (endY - warpY) * (endY - warpY);
//计算 图像的高 宽 通道数量
int height = img.rows;
int width = img.cols;
int chan = img.channels();
auto Abs = [&](float f) {
return f > 0 ? f : -f;
};
for (int i = 0; i < width; i++)
{
for (int j = 0; j < height; j++)
{
// # 计算该点是否在形变圆的范围之内
//# 优化，第一步，直接判断是会在（startX, startY)的矩阵框中
if ((Abs(i - warpX) > radius) && (Abs(j - warpY) > radius))
continue;
float distance = (i - warpX) * (i - warpX) + (j - warpY) * (j - warpY);
if (distance < ddradius)
{
//# 计算出（i, j）坐标的原坐标
//# 计算公式中右边平方号里的部分
float ratio = (ddradius - distance) / (ddradius - distance + mc);
ratio *= ratio;
//映射原位置
float UX = i - ratio * (endX - warpX);
float UY = j - ratio * (endY - warpY);
//根据双线性插值得到UX UY的值
BilinearInsert(img, dst, UX, UY, i, j);
//改变当前的值
}
}
}
return dst;
}
Mat LocalTranslationWarp_Eye(Mat& img, int warpX, int warpY, int endX, int endY, float radius)
{
//平移距离
Mat dst = img.clone();
float ddradius = radius * radius;
//计算|m-c|^2
size_t mc = (endX - warpX) * (endX - warpX) + (endY - warpY) * (endY - warpY);
//计算 图像的高 宽 通道数量
int height = img.rows;
int width = img.cols;
int chan = img.channels();
auto Abs = [&](float f) {
return f > 0 ? f : -f;
};
for (int i = 0; i < width; i++)
{
for (int j = 0; j < height; j++)
{
// # 计算该点是否在形变圆的范围之内
//# 优化，第一步，直接判断是会在（startX, startY)的矩阵框中
if ((Abs(i - warpX) > radius) && (Abs(j - warpY) > radius))
continue;
float distance = (i - warpX) * (i - warpX) + (j - warpY) * (j - warpY);
if (distance < ddradius)
{
float rnorm = sqrt(distance) / radius;
float ratio = 1 - (rnorm - 1) * (rnorm - 1) * 0.5;
//映射原位置
float UX = warpX + ratio * (i - warpX);
float UY = warpY + ratio * (j - warpY);
//根据双线性插值得到UX UY的值
BilinearInsert(img, dst, UX, UY, i, j);
}
}
}
return dst;
}
Mat numpy_uint8_3c_to_cv_mat(py::array_t<unsigned char>& input) {
if (input.ndim() != 3)
throw std::runtime_error("1-channel image must be 2 dims ");
py::buffer_info buf = input.request();
Mat mat(buf.shape[0], buf.shape[1], CV_8UC3, (unsigned char*)buf.ptr);
//imshow("显示读取图片", mat);
return mat;
}
Mat numpy_uint8_1c_to_cv_mat(py::array_t<unsigned char>& input) {
if (input.ndim() != 2)
throw std::runtime_error("1-channel image must be 2 dims ");
py::buffer_info buf = input.request();
Mat mat(buf.shape[0], buf.shape[1], CV_8UC1, (unsigned char*)buf.ptr);
return mat;
}
py::array_t<unsigned char> cv_mat_uint8_1c_to_numpy(cv::Mat& input) {
py::array_t<unsigned char> dst = py::array_t<unsigned char>({ input.rows,input.cols }, input.data);
return dst;
}
py::array_t<unsigned char> cv_mat_uint8_3c_to_numpy(cv::Mat& input) {
py::array_t<unsigned char> dst = py::array_t<unsigned char>({ input.rows,input.cols,3 }, input.data);
return dst;
}
py::array_t<unsigned char> eyebig(py::array_t<unsigned char>& input, int warpX, int warpY, int endX, int endY, float radius) {
Mat img_rgb = numpy_uint8_3c_to_cv_mat(input);
//cout << "input.rows" << img_rgb.rows << "input.cols" << img_rgb.cols << endl;
Mat dst = LocalTranslationWarp_Eye(img_rgb, warpX, warpY, endX, endY, radius);
//imshow("显示读取图片2", dst);
return cv_mat_uint8_3c_to_numpy(dst);
}
py::array_t<unsigned char> facethin(py::array_t<unsigned char>& input, int warpX, int warpY, int endX, int endY, float radius) {
Mat img_rgb = numpy_uint8_3c_to_cv_mat(input);
//cout << "input.rows" << img_rgb.rows << "input.cols" << img_rgb.cols << endl;
Mat dst = LocalTranslationWarp_Face(img_rgb, warpX, warpY, endX, endY, radius);
//imshow("显示读取图片2", dst);
return cv_mat_uint8_3c_to_numpy(dst);
}
py::array_t<unsigned char> test_rgb_to_gray(py::array_t<unsigned char>& input) {
Mat img_rgb = numpy_uint8_3c_to_cv_mat(input);
Mat dst;
cv::cvtColor(img_rgb, dst, cv::COLOR_RGB2GRAY);
return cv_mat_uint8_1c_to_numpy(dst);
}
PYBIND11_MODULE(py2cpp, m) {
m.doc() = "pybind11 example";
m.def("LocalTranslationWarp_Eye", &LocalTranslationWarp_Eye, "LocalTranslationWarp_Eye");
m.def("test_rgb_to_gray", &test_rgb_to_gray,"test_rgb_to_grayt");
m.def("eyebig", &eyebig, "eyebig");
m.def("facethin", &facethin, "facethin");
}
python代码
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import py2cpp as obj
import dlib
import cv2
import numpy as np
import math
import time
predictor_path = 'shape_predictor_68_face_landmarks.dat'
# 使用dlib自带的frontal_face_detector作为我们的特征提取器
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(predictor_path)
def landmark_dec_dlib_fun(img_src):
#转灰度图
img_gray = cv2.cvtColor(img_src, cv2.COLOR_BGR2GRAY)
#cv2.imshow('grayimg',img_gray)
land_marks = []
rects = detector(img_gray, 0)
print(rects)
for i in range(len(rects)):
land_marks_node = np.matrix([[p.x, p.y] for p in predictor(img_gray, rects[i]).parts()])
# for idx,point in enumerate(land_marks_node):
# # 68点坐标
# pos = (point[0,0],point[0,1])
# print(idx,pos)
# # 利用cv2.circle给每个特征点画一个圈，共68个
# cv2.circle(img_src, pos, 5, color=(0, 255, 0))
# # 利用cv2.putText输出1-68
# font = cv2.FONT_HERSHEY_SIMPLEX
# cv2.putText(img_src, str(idx + 1), pos, font, 0.5, (0, 0, 0), 1, cv2.LINE_AA)
land_marks.append(land_marks_node)
return land_marks
'''
方法1： Interactive Image Warping 局部放大算法
'''
def localTranslationWarp1(srcImg, startX, startY, endX, endY, radius):
ddradius = float(radius * radius)
copyImg = np.zeros(srcImg.shape, np.uint8)
copyImg = srcImg.copy()
t1 = time.time()
# 计算公式中的|m-c|^2
ddmc = (endX - startX) * (endX - startX) + (endY - startY) * (endY - startY)
H, W, C = srcImg.shape
for i in range(W):
for j in range(H):
# 计算该点是否在形变圆的范围之内
# 优化，第一步，直接判断是会在（startX,startY)的矩阵框中
if math.fabs(i - startX) > radius and math.fabs(j - startY) > radius:
continue
distance = (i - startX) * (i - startX) + (j - startY) * (j - startY)
if (distance < ddradius):
# 计算出（i,j）坐标的原坐标
# 计算公式中右边平方号里的部分
# ratio = (ddradius - distance) / (ddradius - distance + ddmc)
# ratio = ratio * ratio
rnorm = math.sqrt(distance) / radius
ratio = 1 - (rnorm - 1)*(rnorm - 1)*0.5
# 映射原位置
UX = startX + ratio * (i - startX)
UY = startY + ratio * (j - startY)
# 根据双线性插值法得到UX，UY的值
value = BilinearInsert(srcImg, UX, UY)
# 改变当前 i ，j的值
copyImg[j, i] = value
return copyImg
'''
方法2： Interactive Image Warping 局部平移算法
'''
def localTranslationWarp2(srcImg, startX, startY, endX, endY, radius):
ddradius = float(radius * radius)
copyImg = np.zeros(srcImg.shape, np.uint8)
copyImg = srcImg.copy()
# 计算公式中的|m-c|^2
ddmc = (endX - startX) * (endX - startX) + (endY - startY) * (endY - startY)
H, W, C = srcImg.shape
for i in range(W):
for j in range(H):
# 计算该点是否在形变圆的范围之内
# 优化，第一步，直接判断是会在（startX,startY)的矩阵框中
if math.fabs(i - startX) > radius and math.fabs(j - startY) > radius:
continue
distance = (i - startX) * (i - startX) + (j - startY) * (j - startY)
if (distance < ddradius):
# 计算出（i,j）坐标的原坐标
# 计算公式中右边平方号里的部分
ratio = (ddradius - distance) / (ddradius - distance + ddmc)
ratio = ratio * ratio
# 映射原位置
UX = i - ratio * (endX - startX)
UY = j - ratio * (endY - startY)
# 根据双线性插值法得到UX，UY的值
value = BilinearInsert(srcImg, UX, UY)
# 改变当前 i ，j的值
copyImg[j, i] = value
return copyImg
predictor_path = 'shape_predictor_68_face_landmarks.dat'
# 使用dlib自带的frontal_face_detector作为我们的特征提取器
detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor(predictor_path)
def landmark_dec_dlib_fun(img_src):
#转灰度图
img_gray = cv2.cvtColor(img_src, cv2.COLOR_BGR2GRAY)
#cv2.imshow('grayimg',img_gray)
land_marks = []
rects = detector(img_gray, 0)
#print(rects)
for i in range(len(rects)):
land_marks_node = np.matrix([[p.x, p.y] for p in predictor(img_gray, rects[i]).parts()])
# for idx,point in enumerate(land_marks_node):
# # 68点坐标
# pos = (point[0,0],point[0,1])
# print(idx,pos)
# # 利用cv2.circle给每个特征点画一个圈，共68个
# cv2.circle(img_src, pos, 5, color=(0, 255, 0))
# # 利用cv2.putText输出1-68
# font = cv2.FONT_HERSHEY_SIMPLEX
# cv2.putText(img_src, str(idx + 1), pos, font, 0.5, (0, 0, 0), 1, cv2.LINE_AA)
land_marks.append(land_marks_node)
return land_marks
# 双线性插值法
def BilinearInsert(src, ux, uy):
w, h, c = src.shape
if c == 3:
x1 = int(ux)
x2 = x1 + 1
y1 = int(uy)
y2 = y1 + 1
part1 = src[y1, x1].astype(np.float64) * (float(x2) - ux) * (float(y2) - uy)
part2 = src[y1, x2].astype(np.float64) * (ux - float(x1)) * (float(y2) - uy)
part3 = src[y2, x1].astype(np.float64) * (float(x2) - ux) * (uy - float(y1))
part4 = src[y2, x2].astype(np.float64) * (ux - float(x1)) * (uy - float(y1))
insertValue = part1 + part2 + part3 + part4
return insertValue.astype(np.int8)
src = cv2.imread('1.jpg')
landmarks = landmark_dec_dlib_fun(src)
# if len(landmarks) == 0:
# return
for landmarks_node in landmarks:
# 第4个点左
left_landmark = landmarks_node[37]
# 第6个点左
left_landmark_down = landmarks_node[27]
# 第14个点右
right_landmark = landmarks_node[44]
# 第16个点右
right_landmark_down = landmarks_node[27]
# 第31个点鼻尖
endPt = landmarks_node[30]
def face_big_auto(x):
#68个关键点二维数组
#landmarks = landmark_dec_dlib_fun(src)
if len(landmarks) == 0:
return
w = cv2.getTrackbarPos('Bigeye', 'How big the eyes do you want')
# 计算第4个点到第6个点的距离作为瘦脸距离
# r_left = math.sqrt(
# (left_landmark[0, 0] - left_landmark_down[0, 0]) * (left_landmark[0, 0] - left_landmark_down[0, 0]) +
# (left_landmark[0, 1] - left_landmark_down[0, 1]) * (left_landmark[0, 1] - left_landmark_down[0, 1]))
#
# # 计算第14个点到第16个点的距离作为瘦脸距离
# r_right = math.sqrt(
# (right_landmark[0, 0] - right_landmark_down[0, 0]) * (right_landmark[0, 0] - right_landmark_down[0, 0]) +
# (right_landmark[0, 1] - right_landmark_down[0, 1]) * (right_landmark[0, 1] - right_landmark_down[0, 1]))
r_left = w
r_right = w
# 大左边眼
# big_image = localTranslationWarp1(src, left_landmark[0, 0], left_landmark[0, 1], endPt[0, 0], endPt[0, 1],
# r_left)
big_image = obj.eyebig(src, left_landmark[0, 0], left_landmark[0, 1], endPt[0, 0], endPt[0, 1],
r_left)
# 大右边眼
# big_image = localTranslationWarp1(big_image, right_landmark[0, 0], right_landmark[0, 1], endPt[0, 0],
# endPt[0, 1], r_right)
big_image = obj.eyebig(big_image, right_landmark[0, 0], right_landmark[0, 1], endPt[0, 0],
endPt[0, 1], r_right)
# 显示
cv2.imshow('How big the eyes do you want', big_image)
cv2.waitKey(1)
def face_thin_auto(x):
s = cv2.getTrackbarPos('Thinface', 'How big the eyes do you want')
# 如果未检测到人脸关键点，就不进行瘦脸
if len(landmarks) == 0:
return
for landmarks_node in landmarks:
#第4个点左
left_landmark = landmarks_node[3]
#第6个点左
left_landmark_down = landmarks_node[5]
#第14个点右
right_landmark = landmarks_node[13]
#第16个点右
right_landmark_down = landmarks_node[15]
#第31个点鼻尖
endPt = landmarks_node[30]
# 计算第4个点到第6个点的距离作为瘦脸距离
# r_left = math.sqrt(
# (left_landmark[0, 0] - left_landmark_down[0, 0]) * (left_landmark[0, 0] - left_landmark_down[0, 0]) +
# (left_landmark[0, 1] - left_landmark_down[0, 1]) * (left_landmark[0, 1] - left_landmark_down[0, 1]))
#
# # 计算第14个点到第16个点的距离作为瘦脸距离
# r_right = math.sqrt(
# (right_landmark[0, 0] - right_landmark_down[0, 0]) * (right_landmark[0, 0] - right_landmark_down[0, 0]) +
# (right_landmark[0, 1] - right_landmark_down[0, 1]) * (right_landmark[0, 1] - right_landmark_down[0, 1]))
r_left = s
r_right = s
# 瘦左边脸
# thin_image = localTranslationWarp2(src, left_landmark[0, 0], left_landmark[0, 1], endPt[0, 0], endPt[0, 1],
# r_left)
thin_image = obj.facethin(src, left_landmark[0, 0], left_landmark[0, 1], endPt[0, 0], endPt[0, 1],
r_left)
# 瘦右边脸
# thin_image = localTranslationWarp2(thin_image, right_landmark[0, 0], right_landmark[0, 1], endPt[0, 0],
# endPt[0, 1], r_right)
thin_image = obj.facethin(thin_image, right_landmark[0, 0], right_landmark[0, 1], endPt[0, 0],
endPt[0, 1], r_right)
# 显示
cv2.imshow('How big the eyes do you want', thin_image)
cv2.waitKey(1)
cv2.namedWindow('How big the eyes do you want')
cv2.createTrackbar('Bigeye', 'How big the eyes do you want', 0, 70, face_big_auto)
cv2.createTrackbar('Thinface', 'How big the eyes do you want', 0, 60, face_thin_auto)
cv2.imshow('How big the eyes do you want',src)
cv2.waitKey(0)
# while True:
# #cv2.imshow('How big the eyes do you want', window)
# ch = cv2.waitKey(1) & 0xFF
# if ch == 27:
# break
cv2.destroyAllWindows()
1659
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
