用Python获取摄像头并实时控制人脸 ！Python真的强！

<tt-image data-tteditor-tag="tteditorTag" contenteditable="false" class="syl1560844965023" data-render-status="finished" data-syl-blot="image" style="box-sizing: border-box; cursor: text; color: rgb(34, 34, 34); font-family: "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei", "WenQuanYi Micro Hei", "Helvetica Neue", Arial, sans-serif; font-size: 16px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; white-space: pre-wrap; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; background-color: rgb(255, 255, 255); text-decoration-style: initial; text-decoration-color: initial; display: block;"> image

<input class="pgc-img-caption-ipt" placeholder="图片描述(最多50字)" value="" style="box-sizing: border-box; outline: 0px; color: rgb(102, 102, 102); position: absolute; left: 187.5px; transform: translateX(-50%); padding: 6px 7px; max-width: 100%; width: 375px; text-align: center; cursor: text; font-size: 12px; line-height: 1.5; background-color: rgb(255, 255, 255); background-image: none; border: 0px solid rgb(217, 217, 217); border-radius: 4px; transition: all 0.2s cubic-bezier(0.645, 0.045, 0.355, 1) 0s;"></tt-image>

实现流程

<bi style="box-sizing: border-box; display: block;">从摄像头获取视频流，并转换为一帧一帧的图像，然后将图像信息传递给opencv这个工具库处理，返回灰度图像（就像你使用本地静态图片一样）</bi><bi style="box-sizing: border-box; display: block;">程序启动后，根据监听器信息，使用一个while循环，不断的加载视频图像，然后返回给opencv工具呈现图像信息。</bi><bi style="box-sizing: border-box; display: block;">创建一个键盘事件监听，按下"d"键，则开始执行面部匹配，并进行面具加载（这个过程是动态的，你可以随时移动）。</bi><bi style="box-sizing: border-box; display: block;">面部匹配使用Dlib中的人脸检测算法来查看是否有人脸存在。如果有，它将为每个人脸创建一个结束位置，眼镜和烟卷会移动到那里结束。</bi><bi style="box-sizing: border-box; display: block;">然后我们需要缩放和旋转我们的眼镜以适合每个人的脸。我们将使用从Dlib的68点模型返回的点集来找到眼睛和嘴巴的中心，并为它们之间的空间旋转。</bi><bi style="box-sizing: border-box; display: block;">在我们实时获取眼镜和烟卷的最终位置后，眼镜和烟卷从屏幕顶部进入，开始匹配你的眼镜和嘴巴。</bi><bi style="box-sizing: border-box; display: block;">假如没有人脸，程序会直接返回你的视频信息，不会有面具移动的效果。</bi><bi style="box-sizing: border-box; display: block;">默认一个周期是4秒钟。然后你可以通过"d"键再次检测。</bi><bi style="box-sizing: border-box; display: block;">程序退出使用"q"键。</bi>

这里我将这个功能抽象成一个面具加载服务，请跟随我的代码一窥究竟吧。

• 1.导入对应的工具包

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">from time import sleep Python学习交流群：1004391443
import cv2
import numpy as np
from PIL import Image
from imutils import face_utils, resize
try:
from dlib import get_frontal_face_detector, shape_predictor
except ImportError:
raise
</pre>

• 创建面具加载服务类DynamicStreamMaskService及其对应的初始化属性:

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">class DynamicStreamMaskService(object):
"""
动态黏贴面具服务
"""
def init(self, saved=False):
self.saved = saved # 是否保存图片
self.listener = True # 启动参数
self.video_capture = cv2.VideoCapture(0) # 调用本地摄像头
self.doing = False # 是否进行面部面具
self.speed = 0.1 # 面具移动速度
self.detector = get_frontal_face_detector() # 面部识别器
self.predictor = shape_predictor("shape_predictor_68_face_landmarks.dat") # 面部分析器
self.fps = 4 # 面具存在时间基础时间
self.animation_time = 0 # 动画周期初始值
self.duration = self.fps * 4 # 动画周期最大值
self.fixed_time = 4 # 画图之后，停留时间
self.max_width = 500 # 图像大小
self.deal, self.text, self.cigarette = None, None, None # 面具对象
</pre>

• 按照上面介绍，我们先实现读取视频流转换图片的函数：

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">def read_data(self):
"""
从摄像头获取视频流，并转换为一帧一帧的图像
:return: 返回一帧一帧的图像信息
"""
_, data = self.video_capture.read()
return data
</pre>

• 接下来我们实现人脸定位函数，及眼镜和烟卷的定位:

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">def get_glasses_info(self, face_shape, face_width):
"""
获取当前面部的眼镜信息
:param face_shape:
:param face_width:
:return:
"""
left_eye = face_shape[36:42]
right_eye = face_shape[42:48]
left_eye_center = left_eye.mean(axis=0).astype("int")
right_eye_center = right_eye.mean(axis=0).astype("int")
y = left_eye_center[1] - right_eye_center[1]
x = left_eye_center[0] - right_eye_center[0]
eye_angle = np.rad2deg(np.arctan2(y, x))
deal = self.deal.resize(
(face_width, int(face_width * self.deal.size[1] / self.deal.size[0])),
resample=Image.LANCZOS)
deal = deal.rotate(eye_angle, expand=True)
deal = deal.transpose(Image.FLIP_TOP_BOTTOM)
left_eye_x = left_eye[0, 0] - face_width // 4
left_eye_y = left_eye[0, 1] - face_width // 6
return {"image": deal, "pos": (left_eye_x, left_eye_y)}
def get_cigarette_info(self, face_shape, face_width):
"""
获取当前面部的烟卷信息
:param face_shape:
:param face_width:
:return:
"""
mouth = face_shape[49:68]
mouth_center = mouth.mean(axis=0).astype("int")
cigarette = self.cigarette.resize(
(face_width, int(face_width * self.cigarette.size[1] / self.cigarette.size[0])),
resample=Image.LANCZOS)
x = mouth[0, 0] - face_width + int(16 * face_width / self.cigarette.size[0])
y = mouth_center[1]
return {"image": cigarette, "pos": (x, y)}
def orientation(self, rects, img_gray):
"""
人脸定位
:return:
"""
faces = []
for rect in rects:
face = {}
face_shades_width = rect.right() - rect.left()
predictor_shape = self.predictor(img_gray, rect)
face_shape = face_utils.shape_to_np(predictor_shape)
face['cigarette'] = self.get_cigarette_info(face_shape, face_shades_width)
face['glasses'] = self.get_glasses_info(face_shape, face_shades_width)
faces.append(face)
return faces
</pre>

• 刚才我们提到了键盘监听事件，这里我们实现一下这个函数：

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">def listener_keys(self):
"""
设置键盘监听事件
:return:
"""
key = cv2.waitKey(1) & 0xFF
if key == ord("q"):
self.listener = False
self.console("程序退出")
sleep(1)
self.exit()
if key == ord("d"):
self.doing = not self.doing
</pre>

• 接下来我们来实现加载面具信息的函数：

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">def init_mask(self):
"""
加载面具
:return:
"""
self.console("加载面具...")
self.deal, self.text, self.cigarette = (
Image.open(x) for x in ["images/deals.png", "images/text.png", "images/cigarette.png"]
)
</pre>

• 上面基本的功能都实现了，我们该实现画图函数了，这个函数原理和之前我写的那篇用AI人脸识别技术实现抖音特效实现是一样的，这里我就不赘述了，可以去github或Python中文社区微信公众号查看。

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">def drawing(self, draw_img, faces):
"""
画图
:param draw_img:
:param faces:
:return:
"""
for face in faces:
if self.animation_time < self.duration - self.fixed_time:
current_x = int(face["glasses"]["pos"][0])
current_y = int(face["glasses"]["pos"][1] * self.animation_time / (self.duration - self.fixed_time))
draw_img.paste(face["glasses"]["image"], (current_x, current_y), face["glasses"]["image"])
cigarette_x = int(face["cigarette"]["pos"][0])
cigarette_y = int(face["cigarette"]["pos"][1] * self.animation_time / (self.duration - self.fixed_time))
draw_img.paste(face["cigarette"]["image"], (cigarette_x, cigarette_y),
face["cigarette"]["image"])
else:
draw_img.paste(face["glasses"]["image"], face["glasses"]["pos"], face["glasses"]["image"])
draw_img.paste(face["cigarette"]["image"], face["cigarette"]["pos"], face["cigarette"]["image"])
draw_img.paste(self.text, (75, draw_img.height // 2 + 128), self.text)
</pre>

• 既然是一个服务类，那该有启动与退出函数吧，最后我们来写一下吧。

<bi style="box-sizing: border-box; display: block;">简单介绍一下这个start()函数, 启动后根据初始化监听信息，不断监听视频流，并将流信息通过opencv转换成图像展示出来。</bi><bi style="box-sizing: border-box; display: block;">并且调用按键监听函数，不断的监听你是否按下"d"键进行面具加载，如果监听成功，则进行图像人脸检测，并移动面具，</bi><bi style="box-sizing: border-box; display: block;">并持续一个周期的时间结束，面具此时会根据你的面部移动而移动。最终呈现文章顶部图片的效果.</bi>

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">def start(self):
"""
启动程序
:return:
"""
self.console("程序启动成功.")
self.init_mask()
while self.listener:
frame = self.read_data()
frame = resize(frame, width=self.max_width)
img_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
rects = self.detector(img_gray, 0)
faces = self.orientation(rects, img_gray)
draw_img = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
if self.doing:
self.drawing(draw_img, faces)
self.animation_time += self.speed
self.save_data(draw_img)
if self.animation_time > self.duration:
self.doing = False
self.animation_time = 0
else:
frame = cv2.cvtColor(np.asarray(draw_img), cv2.COLOR_RGB2BGR)
cv2.imshow("hello mask", frame)
self.listener_keys()
def exit(self):
"""
程序退出
:return:
"""
self.video_capture.release()
cv2.destroyAllWindows()
</pre>

• 最后，让我们试试：

<pre spellcheck="false" style="box-sizing: border-box; margin: 5px 0px; padding: 5px 10px; border: 0px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-variant-numeric: inherit; font-variant-east-asian: inherit; font-weight: 400; font-stretch: inherit; font-size: 16px; line-height: inherit; font-family: inherit; vertical-align: baseline; cursor: text; counter-reset: list-1 0 list-2 0 list-3 0 list-4 0 list-5 0 list-6 0 list-7 0 list-8 0 list-9 0; background-color: rgb(240, 240, 240); border-radius: 3px; white-space: pre-wrap; color: rgb(34, 34, 34); letter-spacing: normal; orphans: 2; text-align: left; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">if name == 'main':
ms = DynamicStreamMaskService()
ms.start()
</pre>

• 写到这里，这个小功能就已经实现了，大家不妨事来使用一下。

<tt-image data-tteditor-tag="tteditorTag" contenteditable="false" class="syl1560844976662 ql-align-center" data-render-status="finished" data-syl-blot="image" style="box-sizing: border-box; cursor: text; text-align: left; color: rgb(34, 34, 34); font-family: "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei", "WenQuanYi Micro Hei", "Helvetica Neue", Arial, sans-serif; font-size: 16px; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: 400; letter-spacing: normal; orphans: 2; text-indent: 0px; text-transform: none; white-space: pre-wrap; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; background-color: rgb(255, 255, 255); text-decoration-style: initial; text-decoration-color: initial; display: block;"> image

<input class="pgc-img-caption-ipt" placeholder="图片描述(最多50字)" value="" style="box-sizing: border-box; outline: 0px; color: rgb(102, 102, 102); position: absolute; left: 187.5px; transform: translateX(-50%); padding: 6px 7px; max-width: 100%; width: 375px; text-align: center; cursor: text; font-size: 12px; line-height: 1.5; background-color: rgb(255, 255, 255); background-image: none; border: 0px solid rgb(217, 217, 217); border-radius: 4px; transition: all 0.2s cubic-bezier(0.645, 0.045, 0.355, 1) 0s;"></tt-image>

写在最后

• 关注 https://github.com/tomoncle/face-detection-induction-course 这个仓库，不定时更新更多的人脸识别小示例。