当前位置：首页 > ds >正文

PiscCode使用MediaPipe Face Landmarker实现实时人脸特征点检测

ds 2025/8/21 14:50:53

概述

人脸特征点检测是计算机视觉领域的重要技术，广泛应用于人脸识别、表情分析、虚拟化妆和美颜滤镜等场景。本文将介绍如何使用MediaPipe的Face Landmarker模型实现实时人脸特征点检测，并在视频帧上绘制人脸网格、轮廓和虹膜特征。

MediaPipe Face Landmarker模型

MediaPipe Face Landmarker是一个轻量级且高效的人脸特征点检测模型，能够检测468个3D人脸特征点，包括:

人脸网格 (FACEMESH_TESSELATION) - 覆盖整个面部的密集网格
人脸轮廓 (FACEMESH_CONTOURS) - 定义人脸外部和内部特征的轮廓线
虹膜特征 (FACEMESH_IRISES) - 精确的眼睛虹膜位置

该模型还支持输出面部混合形状参数和面部变换矩阵，可用于高级的面部分析和3D重建任务。

代码实现解析

1. 初始化FaceLandmarker

def __init__(self, model_path="path/to/face_landmarker.task"):base_options = python.BaseOptions(model_asset_path=model_path)options = vision.FaceLandmarkerOptions(base_options=base_options,output_face_blendshapes=True,output_facial_transformation_matrixes=True,num_faces=1  # 设置检测的人脸数量)self.detector = vision.FaceLandmarker.create_from_options(options)

2. 绘制人脸特征点

def _draw_landmarks_on_image(self, rgb_image, detection_result):# 创建图像副本并绘制三类特征点连接solutions.drawing_utils.draw_landmarks(image=annotated_image,landmark_list=face_landmarks_proto,connections=mp.solutions.face_mesh.FACEMESH_TESSELATION,  # 人脸网格# ...其他参数)# 绘制轮廓和虹膜特征点...

3. 处理视频帧

import cv2
import numpy as np
import mediapipe as mp
from mediapipe import solutions
from mediapipe.framework.formats import landmark_pb2
from mediapipe.tasks import python
from mediapipe.tasks.python import visionclass FrameObject:def __init__(self, model_path="地址/face_landmarker.task"):"""初始化 Mediapipe FaceLandmarker"""base_options = python.BaseOptions(model_asset_path=model_path)options = vision.FaceLandmarkerOptions(base_options=base_options,output_face_blendshapes=True,output_facial_transformation_matrixes=True,num_faces=1)self.detector = vision.FaceLandmarker.create_from_options(options)def _draw_landmarks_on_image(self, rgb_image, detection_result):"""在图像上绘制人脸网格、轮廓和虹膜"""face_landmarks_list = detection_result.face_landmarksannotated_image = np.copy(rgb_image)for face_landmarks in face_landmarks_list:face_landmarks_proto = landmark_pb2.NormalizedLandmarkList()face_landmarks_proto.landmark.extend([landmark_pb2.NormalizedLandmark(x=lm.x, y=lm.y, z=lm.z)for lm in face_landmarks])# 绘制三类连接solutions.drawing_utils.draw_landmarks(image=annotated_image,landmark_list=face_landmarks_proto,connections=mp.solutions.face_mesh.FACEMESH_TESSELATION,landmark_drawing_spec=None,connection_drawing_spec=mp.solutions.drawing_styles.get_default_face_mesh_tesselation_style())solutions.drawing_utils.draw_landmarks(image=annotated_image,landmark_list=face_landmarks_proto,connections=mp.solutions.face_mesh.FACEMESH_CONTOURS,landmark_drawing_spec=None,connection_drawing_spec=mp.solutions.drawing_styles.get_default_face_mesh_contours_style())solutions.drawing_utils.draw_landmarks(image=annotated_image,landmark_list=face_landmarks_proto,connections=mp.solutions.face_mesh.FACEMESH_IRISES,landmark_drawing_spec=None,connection_drawing_spec=mp.solutions.drawing_styles.get_default_face_mesh_iris_connections_style())return annotated_imagedef do(self, frame, device=None):"""处理单帧图像，返回绘制人脸网格后的帧"""if frame is None:return Nonemp_image = mp.Image(image_format=mp.ImageFormat.SRGB,data=cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))detection_result = self.detector.detect(mp_image)annotated = self._draw_landmarks_on_image(mp_image.numpy_view(), detection_result)return cv2.cvtColor(annotated, cv2.COLOR_RGB2BGR)