来自云龙湖轮廓分明的月亮 发表于 2025-3-7 21:50:19

计算机视觉之dlib人脸关键点绘制及微笑测试

dlib人脸关键点绘制及微笑测试



1 dlib人脸关键点

1.1 dlib

dlib 是一个强盛的呆板学习库,广泛用于人脸检测和人脸关键点检测。它提供了一个预训练的 68 点人脸关键点检测模型,可以正确地定位人脸的各个部位(如眼睛、鼻子、嘴巴等)
1.2 人脸关键点检测

dlib 的 68 点人脸关键点检测模型基于 HOG(Histogram of Oriented Gradients)特性和线性分类器,联合了形状预测算法。它可以检测人脸的以下区域:
下巴(0-16)
右眉毛(17-21)
左眉毛(22-26)
鼻子(27-35)
右眼(36-41)
左眼(42-47)
嘴巴(48-67)
https://i-blog.csdnimg.cn/direct/19d815f74b4d4255a772147a4e4c6db1.png
1.3 检测模型

dlib 提供了一个预训练的 68 点人脸关键点检测模型,可以从以下链接下载:
https://github.com/davisking/dlib-models/blob/master/shape_predictor_68_face_landmarks.dat.bz2/
1.4 凸包

凸包(Convex Hull) 是计算多少中的一个重要概念,指的是在二维或更高维空间中,包含一组点的最小凸多边形或凸多面体。凸包在图像处置惩罚、计算机视觉、模式辨认等领域有广泛应用,例如在人脸关键点检测中,可以用凸包来界说人脸区域的边界。
1.5 笑容检测

界说了两个函数,MAR:衡量嘴巴的伸开水平,
和MJR:衡量嘴巴宽度与下巴宽度的比例,
人脸关键点如上,当微笑时嘴巴长款和面颊长度都会发生改变,通过两个函数举行比较检测,举行判定是否微笑
def MAR(shape):
    x = shape
    y = shape.reshape(1,2)
    A = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    B = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    C = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    D = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))

    return ((A+B+C)/3)/D

def MJR(shape):
    M = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    J = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    return M/J
1.6 函数



[*]detector = dlib.get_frontal_face_detector() ,加载人脸检测器
[*]predictor = dlib.shape_predictor(‘shape_predictor_68_face_landmarks.dat’) 关键点预测器
[*]detector(gray, 1) ,检测人脸

[*]gray检测的灰度图
[*]1 表示对图像举行上采样次数

2 人脸检测代码

2.1 关键点绘制

代码展示:
import cv2
import numpy as np
import dlib

img = cv2.imread('lyf.png')
detector = dlib.get_frontal_face_detector()
faces = detector(img,0)
predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
for face in faces:
    shape = predictor(img,face)
    landmarks = np.array([ for p in shape.parts()])
    for idx,point in enumerate(landmarks):
      pos = ,point]
      cv2.circle(img,pos,2,color=(0,255,0),thickness=-1)
      cv2.putText(img,str(idx),pos,cv2.FONT_HERSHEY_SCRIPT_SIMPLEX,0.4,(255,255,255),1,cv2.LINE_AA)
cv2.imshow('img',img)
cv2.waitKey(0)
cv2.destroyAllWindows()
运行结果:
https://i-blog.csdnimg.cn/direct/54628d4c9ea9444ca86ba7af3d17ed66.png
2.2 关键点连线

代码展示:
import cv2
import numpy as np
import dlib

def drawLine(start,end):
    pts = shape
    for l inrange(1,len(pts)):
      pta = tuple(pts)
      ptb = tuple(pts)
      cv2.line(img,pta,ptb,(0,255,0),1)

def drawConvexHull(start,end):
    facial = shape
    mouthHull = cv2.convexHull(facial)
    cv2.drawContours(img,,-1,(0,255,0),1)

img = cv2.imread('lyf.png')
detector = dlib.get_frontal_face_detector()
faces = detector(img,0)
predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')
for face in faces:
    shape = predictor(img,face)
    shape = np.array([ for p in shape.parts()])
    drawConvexHull(36,41)
    drawConvexHull(42,47)
    drawConvexHull(48, 59)
    drawConvexHull(60, 67)

    drawLine(0,17)
    drawLine(17, 22)
    drawLine(22, 27)
    drawLine(27, 36)

cv2.imshow('img',img)
cv2.waitKey(0)
cv2.destroyAllWindows()
运行结果:
https://i-blog.csdnimg.cn/direct/5168faa913084c608c61cb14ce49b8d2.png
2.3 微笑检测

代码展示:
import cv2import numpy as npimport dlibdetector = dlib.get_frontal_face_detector()predictor = dlib.shape_predictor('shape_predictor_68_face_landmarks.dat')v = cv2.VideoCapture('jjy_dyx.mp4')from sklearn.metrics.pairwise import euclidean_distancesfrom PIL import Image, ImageDraw, ImageFontdef cv2AddChineseText(img, text, position, textColor=(255, 255, 255), textSize=30):    """ 向图片中添加中文 """    if (isinstance(img, np.ndarray)):# 判定是否OpenCV图片类型      img = Image.fromarray(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))#实现array到image的转换    draw = ImageDraw.Draw(img)# 在img图片上创建一个画图的对象    # 字体的格式    fontStyle = ImageFont.truetype("simsun.ttc", textSize, encoding="utf-8")    draw.text(position, text, textColor, font=fontStyle) # 绘制文本    return cv2.cvtColor(np.asarray(img), cv2.COLOR_RGB2BGR)# 转换回OpenCV格式def MAR(shape):
    x = shape
    y = shape.reshape(1,2)
    A = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    B = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    C = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    D = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))

    return ((A+B+C)/3)/D

def MJR(shape):
    M = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    J = euclidean_distances(shape.reshape(1,2),shape.reshape(1,2))
    return M/J
while True:    r,img = v.read()    if not r:      break    faces = detector(img,0)    for face in faces:      shape = predictor(img,face)      shape= np.array([ for p in shape.parts()])      mar = MAR(shape)      mjr =MJR(shape)      result = '正常'      print('mar:',mar,'mjr:',mjr)      if mar>0.5:            result = '大笑'      elif mjr>0.4:            result = '微笑'      mouthHull = cv2.convexHull(shape)      img = cv2AddChineseText(img,result,mouthHull,1)      cv2.drawContours(img,,-1,(0,255,0),1)    cv2.imshow('img', img)    key = cv2.waitKey(1)    if key == 32:      breakv.release()cv2.waitKey(0)cv2.destroyAllWindows() 运行结果:
https://i-blog.csdnimg.cn/direct/5973ad617d2a4701b31c6426b6128c27.png

免责声明:如果侵犯了您的权益,请联系站长,我们会及时删除侵权内容,谢谢合作!更多信息从访问主页:qidao123.com:ToB企服之家,中国第一个企服评测及商务社交产业平台。
页: [1]
查看完整版本: 计算机视觉之dlib人脸关键点绘制及微笑测试