LBPH人脸识别简介

LBPH（Local Binary Patterns Histograms）是一种高效的人脸识别算法，它结合了局部二值模式(LBP)和直方图技术。LBPH的主要优势在于对光照变化不敏感，计算效率高，适用于实时人脸识别系统。

LBPH人脸识别实现步骤

Python代码实现

1. 安装依赖库

pip install opencv-python
pip install numpy

2. 准备数据集

创建如下目录结构，每个人一个文件夹：

dataset/
├── person1/
│   ├── 1.jpg
│   ├── 2.jpg
│   └── ...
├── person2/
│   ├── 1.jpg
│   ├── 2.jpg
│   └── ...
└── ...

3. 完整人脸识别代码

import cv2
import os
import numpy as np

# 加载人脸检测器
face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml')

def detect_faces(img):
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5)
    return faces

def prepare_training_data(data_folder_path):
    dirs = os.listdir(data_folder_path)
    faces = []
    labels = []
    label_dict = {}
    current_label = 0
    
    for dir_name in dirs:
        if not dir_name.startswith("."):
            label_dict[current_label] = dir_name
            subject_dir_path = os.path.join(data_folder_path, dir_name)
            subject_images_names = os.listdir(subject_dir_path)
            
            for image_name in subject_images_names:
                if image_name.startswith("."):
                    continue
                    
                image_path = os.path.join(subject_dir_path, image_name)
                image = cv2.imread(image_path)
                
                # 检测人脸
                faces_rect = detect_faces(image)
                if len(faces_rect) != 1:
                    continue
                    
                (x, y, w, h) = faces_rect[0]
                face = image[y:y+w, x:x+h]
                face = cv2.resize(face, (200, 200))
                face = cv2.cvtColor(face, cv2.COLOR_BGR2GRAY)
                
                faces.append(face)
                labels.append(current_label)
            
            current_label += 1
            
    return faces, labels, label_dict

# 准备训练数据
data_path = "dataset"
faces, labels, label_dict = prepare_training_data(data_path)

# 创建LBPH人脸识别器
recognizer = cv2.face.LBPHFaceRecognizer_create()

# 训练模型
recognizer.train(faces, np.array(labels))

# 保存模型
recognizer.save("lbph_model.yml")
np.save("label_dict.npy", label_dict)

# 测试识别
def predict(test_img):
    img = test_img.copy()
    faces_rect = detect_faces(img)
    
    for (x, y, w, h) in faces_rect:
        face = img[y:y+w, x:x+h]
        face = cv2.resize(face, (200, 200))
        gray_face = cv2.cvtColor(face, cv2.COLOR_BGR2GRAY)
        
        label, confidence = recognizer.predict(gray_face)
        
        # 获取标签对应的名称
        label_text = label_dict[label]
        
        # 绘制结果
        cv2.rectangle(img, (x, y), (x+w, y+h), (0, 255, 0), 2)
        cv2.putText(img, f"{label_text} ({confidence:.2f})", (x, y-10), 
                   cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 0), 2)
    
    return img

# 加载测试图像
test_img = cv2.imread("test.jpg")
predicted_img = predict(test_img)

# 显示结果
cv2.imshow("Prediction", predicted_img)
cv2.waitKey(0)
cv2.destroyAllWindows()

LBPH参数调优指南

LBPH人脸识别应用场景