做完了，但是是大粪
This commit is contained in:
@@ -0,0 +1,78 @@
|
||||
import cv2 as cv
|
||||
import numpy as np
|
||||
import dlib
|
||||
import joblib
|
||||
|
||||
|
||||
# 画框函数
|
||||
def draw_rect(img, faces):
|
||||
for face in faces:
|
||||
cv.rectangle(img, (face.left(), face.top()), (face.right(), face.bottom()), (0, 0, 255), 32)
|
||||
return img
|
||||
|
||||
|
||||
# 检测人脸并画框
|
||||
def detect_face(img):
|
||||
img_gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY)
|
||||
result = detector(img_gray, 1)
|
||||
img_result = draw_rect(img, result)
|
||||
return img_result
|
||||
|
||||
|
||||
# 提取人脸特征
|
||||
def extract_face_feature(img):
|
||||
faces = detector(img)
|
||||
if len(faces) > 0:
|
||||
face = faces[0]
|
||||
landmarks = predictor(img, face)
|
||||
face_descriptor = face_rec_model.compute_face_descriptor(img, landmarks)
|
||||
return np.array(face_descriptor)
|
||||
else:
|
||||
return None
|
||||
|
||||
|
||||
# 加载分类器
|
||||
classifier = joblib.load('models/classifier.pkl')
|
||||
|
||||
# 读取测试图片
|
||||
positive_sample = cv.imread('cache/positive_test.jpg')
|
||||
positive_sample_rgb = cv.cvtColor(positive_sample, cv.COLOR_BGR2RGB)
|
||||
|
||||
negative_sample = cv.imread('cache/negative_test.jpg')
|
||||
negative_sample_rgb = cv.cvtColor(negative_sample, cv.COLOR_BGR2RGB)
|
||||
|
||||
# 加载dlib人脸检测模型
|
||||
detector = dlib.get_frontal_face_detector()
|
||||
predictor = dlib.shape_predictor("models/shape_predictor_68_face_landmarks_GTX.dat")
|
||||
face_rec_model = dlib.face_recognition_model_v1('models/dlib_face_recognition_resnet_model_v1.dat')
|
||||
|
||||
# 检测人脸
|
||||
positive_result = detect_face(positive_sample)
|
||||
negative_result = detect_face(negative_sample)
|
||||
|
||||
# 计算人脸特征
|
||||
positive_feature = extract_face_feature(positive_sample_rgb)
|
||||
positive_feature = np.array(positive_feature).reshape(1, -1)
|
||||
|
||||
negative_feature = extract_face_feature(negative_sample_rgb)
|
||||
negative_feature = np.array(negative_feature).reshape(1, -1)
|
||||
|
||||
# 通过分类器预测
|
||||
positive_predict = classifier.predict(positive_feature)
|
||||
negative_predict = classifier.predict(negative_feature)
|
||||
|
||||
# 展示
|
||||
positive_result_resized = cv.resize(positive_result,
|
||||
(int(positive_result.shape[1] * 0.1), int(positive_result.shape[0] * 0.1)))
|
||||
positive_result_resized = cv.putText(positive_result_resized, str(positive_predict), (50, 50), cv.FONT_HERSHEY_SIMPLEX,
|
||||
1, (0, 0, 255), 2)
|
||||
|
||||
negative_result_resized = cv.resize(negative_result,
|
||||
(int(negative_result.shape[1] * 0.1), int(negative_result.shape[0] * 0.1)))
|
||||
negative_result_resized = cv.putText(negative_result_resized, str(negative_predict), (50, 50), cv.FONT_HERSHEY_SIMPLEX,
|
||||
1, (0, 0, 255), 2)
|
||||
|
||||
cv.imshow('positive_result', positive_result_resized)
|
||||
cv.imshow('negative_result', negative_result_resized)
|
||||
cv.waitKey(0)
|
||||
cv.destroyAllWindows()
|
||||
|
||||
+47
-41
@@ -8,88 +8,94 @@ from sklearn.model_selection import train_test_split
|
||||
from sklearn.svm import SVC
|
||||
from tqdm import tqdm
|
||||
|
||||
|
||||
# 方法区
|
||||
## 转换为二值图和GRB图
|
||||
## 转换为二值图和RGB图
|
||||
def img_cvt(img):
|
||||
img_gray = cv.cvtColor(img, cv2.COLOR_BGR2GRAY)
|
||||
_,img_gray=cv.threshold(img_gray,127,255,cv.THRESH_BINARY)
|
||||
img_gray = cv.blur(img_gray,(3,3))
|
||||
_, img_gray = cv.threshold(img_gray, 127, 255, cv.THRESH_BINARY)
|
||||
img_gray = cv.blur(img_gray, (3, 3))
|
||||
img_rgb = cv.cvtColor(img, cv.COLOR_BGR2RGB)
|
||||
return img_gray,img_rgb
|
||||
return img_gray, img_rgb
|
||||
|
||||
|
||||
## 提取人脸特征
|
||||
def extract_face_feature(img):
|
||||
faces=detector(img)
|
||||
if len(faces)>0:
|
||||
face=faces[0]
|
||||
landmarks=predictor(img,face)
|
||||
face_descriptor=face_rec_model.compute_face_descriptor(img,landmarks)
|
||||
faces = detector(img)
|
||||
if len(faces) > 0:
|
||||
face = faces[0]
|
||||
landmarks = predictor(img, face)
|
||||
face_descriptor = face_rec_model.compute_face_descriptor(img, landmarks)
|
||||
return np.array(face_descriptor)
|
||||
else:
|
||||
return None
|
||||
|
||||
|
||||
## 生成伪样品
|
||||
def gen_negative_samples(features_dim=128):
|
||||
return np.random.uniform(-1,1,features_dim)
|
||||
return np.random.uniform(-1, 1, features_dim)
|
||||
|
||||
|
||||
# 变量区
|
||||
dataset_video=cv.VideoCapture('cache/dataset.mp4')
|
||||
datasets=[]
|
||||
datasets_rgb=[]
|
||||
features=[]
|
||||
negative_samples=[]
|
||||
dataset_video = cv.VideoCapture('cache/dataset.mp4')
|
||||
datasets = []
|
||||
datasets_rgb = []
|
||||
features = []
|
||||
negative_samples = []
|
||||
|
||||
# 收集数据集
|
||||
## 采用视频的格式收集图片
|
||||
while(True):
|
||||
ret,frame=dataset_video.read()
|
||||
while (True):
|
||||
ret, frame = dataset_video.read()
|
||||
if not ret:
|
||||
break
|
||||
datasets.append(frame)
|
||||
|
||||
# 加载人脸检测模型
|
||||
detector=dlib.get_frontal_face_detector()
|
||||
predictor=dlib.shape_predictor("models/shape_predictor_68_face_landmarks_GTX.dat")
|
||||
face_rec_model=dlib.face_recognition_model_v1('models/dlib_face_recognition_resnet_model_v1.dat')
|
||||
detector = dlib.get_frontal_face_detector()
|
||||
predictor = dlib.shape_predictor("models/shape_predictor_68_face_landmarks_GTX.dat")
|
||||
face_rec_model = dlib.face_recognition_model_v1('models/dlib_face_recognition_resnet_model_v1.dat')
|
||||
|
||||
# 预处理数据集
|
||||
for dataset in tqdm(datasets,desc='正在预处理数据集'):
|
||||
_,dataset_rgb = img_cvt(dataset)
|
||||
for dataset in tqdm(datasets, desc='正在预处理数据集'):
|
||||
_, dataset_rgb = img_cvt(dataset)
|
||||
datasets_rgb.append(dataset_rgb)
|
||||
|
||||
# 求人脸特征
|
||||
for dataset in tqdm(datasets_rgb,desc='正在求人脸特征'):
|
||||
face_descriptor=extract_face_feature(dataset)
|
||||
for dataset in tqdm(datasets_rgb, desc='正在求人脸特征'):
|
||||
face_descriptor = extract_face_feature(dataset)
|
||||
if face_descriptor is not None:
|
||||
features.append(face_descriptor)
|
||||
|
||||
# 生成负样本
|
||||
for i in tqdm(range(len(features)*15),desc='正在生成负样本'):
|
||||
negative_sample=gen_negative_samples()
|
||||
for i in tqdm(range(len(features) * 2048), desc='正在生成负样本'):
|
||||
negative_sample = gen_negative_samples()
|
||||
negative_samples.append(negative_sample)
|
||||
|
||||
# 编码数据集
|
||||
positive_labels=[1]*len(features)
|
||||
negative_labels=[0]*len(negative_samples)
|
||||
positive_labels = [1] * len(features)
|
||||
negative_labels = [0] * len(negative_samples)
|
||||
|
||||
samples=features+negative_samples
|
||||
labels=positive_labels+negative_labels
|
||||
samples = features + negative_samples
|
||||
labels = positive_labels + negative_labels
|
||||
|
||||
X_train=np.array(samples)
|
||||
Y_train=np.array(labels)
|
||||
X_train = np.array(samples)
|
||||
Y_train = np.array(labels)
|
||||
|
||||
# 分离数据集和训练集
|
||||
X_train,X_test,Y_train,Y_test=train_test_split(X_train,Y_train,test_size=0.2,random_state=42,stratify=Y_train)
|
||||
X_train, X_test, Y_train, Y_test = train_test_split(X_train, Y_train, test_size=0.2, random_state=42, stratify=Y_train)
|
||||
|
||||
# 训练
|
||||
print('正在训练分类器')
|
||||
classifier=SVC(kernel='linear',C=1.0,random_state=42)
|
||||
classifier.fit(X_train,Y_train)
|
||||
classifier = SVC(kernel='linear', C=1.0, random_state=42)
|
||||
classifier.fit(X_train, Y_train)
|
||||
|
||||
# 评估分类器性能
|
||||
Y_pred=[]
|
||||
for test_sample in tqdm(X_test,desc='评估分类器'):
|
||||
Y_pred.append(classifier.predict(test_sample.reshape(1,-1)))
|
||||
accuracy=accuracy_score(Y_test,Y_pred)
|
||||
print(f'准确率：{accuracy*100:.2f}%')
|
||||
Y_pred = []
|
||||
for test_sample in tqdm(X_test, desc='评估分类器'):
|
||||
Y_pred.append(classifier.predict(test_sample.reshape(1, -1)))
|
||||
accuracy = accuracy_score(Y_test, Y_pred)
|
||||
print(f'准确率：{accuracy * 100:.5f}%')
|
||||
|
||||
# 保存模型
|
||||
joblib.dump(classifier,'models/classifier.pkl')
|
||||
joblib.dump(classifier, 'models/classifier.pkl')
|
||||
|
||||
Reference in New Issue
Block a user