使用OpenCV和Dlib进行人脸特征提取 基于人工智能的课堂人脸考勤系统 使用深度学习模型进行活体检测和专注度检测 使用Qt Designer设计界面基于人工智能的课堂人脸考勤系统文章目录使用OpenCV和Dlib进行人脸特征提取 基于人工智能的课堂人脸考勤系统 使用深度学习模型进行活体检测和专注度检测 使用Qt Designer设计界面1. 环境搭建安装依赖2. 人脸库录入录入人脸库代码3. 人脸识别人脸识别代码4. 活体检测和专注度检测活体检测代码专注度检测代码5. UI设计与结果展示设计界面主程序代码系统主要分为三个部分第一个部分人脸库的录入通过录入人脸裤记录学生信息。第二部分人脸识别选择班级考勤时间后打开摄像头进行人脸考勤。第三部分活体检测和专注度检测判断学生是否在听课以及听课的专注度。1基于人工智能的课堂人脸考勤系统人脸库录入、人脸识别、活体检测和专注度检测。关键代码示例1. 环境搭建确保你的开发环境已经安装了必要的库和工具。安装依赖# 创建并激活虚拟环境conda create-nface_attendancepython3.8conda activate face_attendance# 安装相关库pipinstallopencv-python numpy dlib imutils PyQt5 tensorflow keras scikit-learn2. 人脸库录入使用OpenCV和Dlib进行人脸特征提取并将特征存储在数据库中。录入人脸库代码importcv2importosimportnumpyasnpimportdlibfromimutilsimportface_utilsimportsqlite3# 初始化Dlib的人脸检测器和特征提取器detectordlib.get_frontal_face_detector()predictordlib.shape_predictor(shape_predictor_68_face_landmarks.dat)face_recogdlib.face_recognition_model_v1(dlib_face_recognition_resnet_model_v1.dat)defadd_face_to_db(name,image_path):connsqlite3.connect(faces.db)cconn.cursor()# 创建表如果不存在c.execute(CREATE TABLE IF NOT EXISTS faces (name TEXT, encoding BLOB))# 加载图像并转换为灰度图imagecv2.imread(image_path)graycv2.cvtColor(image,cv2.COLOR_BGR2GRAY)# 检测人脸rectsdetector(gray,0)iflen(rects)0:print(No face detected in the image.)returnforrectinrects:shapepredictor(gray,rect)face_descriptorface_recog.compute_face_descriptor(image,shape)# 将特征向量转换为二进制格式并存储到数据库中face_descriptor_binnp.array(face_descriptor).tobytes()c.execute(INSERT INTO faces (name, encoding) VALUES (?, ?),(name,face_descriptor_bin))conn.commit()conn.close()# 示例添加一张人脸到数据库add_face_to_db(Student1,path/to/student1.jpg)3. 人脸识别使用摄像头实时捕捉图像并与数据库中的人脸特征进行比对。人脸识别代码importcv2importnumpyasnpimportdlibimportsqlite3fromimutilsimportface_utils# 初始化Dlib的人脸检测器和特征提取器detectordlib.get_frontal_face_detector()predictordlib.shape_predictor(shape_predictor_68_face_landmarks.dat)face_recogdlib.face_recognition_model_v1(dlib_face_recognition_resnet_model_v1.dat)defload_faces_from_db():connsqlite3.connect(faces.db)cconn.cursor()c.execute(SELECT name, encoding FROM faces)rowsc.fetchall()known_names[]known_encodings[]forrowinrows:namerow[0]encodingnp.frombuffer(row[1],dtypenp.float64)known_names.append(name)known_encodings.append(encoding)conn.close()returnknown_names,known_encodingsdefrecognize_faces(frame,known_names,known_encodings):graycv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)rectsdetector(gray,0)forrectinrects:shapepredictor(gray,rect)face_descriptorface_recog.compute_face_descriptor(frame,shape)matchesdlib.compare_face_descriptors(known_encodings,face_descriptor,tolerance0.6)ifTrueinmatches:match_indexmatches.index(True)nameknown_names[match_index]print(fRecognized:{name})returnnamereturnNone# 实时人脸识别defreal_time_face_recognition():known_names,known_encodingsload_faces_from_db()capcv2.VideoCapture(0)whileTrue:ret,framecap.read()ifnotret:breakrecognized_namerecognize_faces(frame,known_names,known_encodings)ifrecognized_name:cv2.putText(frame,recognized_name,(10,30),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,0),2)cv2.imshow(Face Recognition,frame)ifcv2.waitKey(1)0xFFord(q):breakcap.release()cv2.destroyAllWindows()# 运行实时人脸识别real_time_face_recognition()4. 活体检测和专注度检测使用深度学习模型进行活体检测和专注度检测。活体检测代码importcv2importnumpyasnpfromkeras.modelsimportload_model# 加载预训练的活体检测模型modelload_model(liveness.model)defis_living_face(frame):face_cascadecv2.CascadeClassifier(haarcascade_frontalface_default.xml)graycv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)facesface_cascade.detectMultiScale(gray,scaleFactor1.1,minNeighbors5,minSize(30,30))iflen(faces)0:(x,y,w,h)faces[0]face_roiframe[y:yh,x:xw]face_roicv2.resize(face_roi,(128,128))face_roiface_roi.astype(float)/255.0face_roinp.expand_dims(face_roi,axis0)predmodel.predict(face_roi)[0]ifpred[1]pred[0]:returnTruereturnFalse# 示例检查当前帧是否为活体人脸capcv2.VideoCapture(0)whileTrue:ret,framecap.read()ifnotret:breakifis_living_face(frame):print(Living face detected!)else:print(Not a living face.)cv2.imshow(Liveness Detection,frame)ifcv2.waitKey(1)0xFFord(q):breakcap.release()cv2.destroyAllWindows()专注度检测代码importcv2importnumpyasnpfromkeras.modelsimportload_model# 加载预训练的专注度检测模型modelload_model(attention.model)defget_attention_level(frame):face_cascadecv2.CascadeClassifier(haarcascade_frontalface_default.xml)graycv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)facesface_cascade.detectMultiScale(gray,scaleFactor1.1,minNeighbors5,minSize(30,30))iflen(faces)0:(x,y,w,h)faces[0]face_roiframe[y:yh,x:xw]face_roicv2.resize(face_roi,(128,128))face_roiface_roi.astype(float)/255.0face_roinp.expand_dims(face_roi,axis0)predmodel.predict(face_roi)[0]attention_levelnp.argmax(pred)returnattention_levelreturnNone# 示例获取当前帧的专注度水平capcv2.VideoCapture(0)whileTrue:ret,framecap.read()ifnotret:breakattention_levelget_attention_level(frame)ifattention_levelisnotNone:print(fAttention Level:{attention_level})cv2.imshow(Attention Detection,frame)ifcv2.waitKey(1)0xFFord(q):breakcap.release()cv2.destroyAllWindows()5. UI设计与结果展示使用PyQt5设计用户界面并展示检测和跟踪结果。设计界面使用Qt Designer设计界面并导出为.ui文件然后使用pyuic5转换为Python代码。pyuic5-xyour_design.ui-oui_design.py主程序代码importsysfromPyQt5.QtWidgetsimportQApplication,QMainWindow,QFileDialogfromPyQt5.QtGuiimportQImage,QPixmapfromPyQt5.QtCoreimportQTimerimportcv2fromui_designimportUi_MainWindowclassFaceAttendanceApp(QMainWindow,Ui_MainWindow):def__init__(self):super().__init__()self.setupUi(self)self.timerQTimer()self.timer.timeout.connect(self.update_frame)self.capNoneself.known_names,self.known_encodingsload_faces_from_db()defstart_camera(self):self.capcv2.VideoCapture(0)self.timer.start(30)defupdate_frame(self):ret,frameself.cap.read()ifret:recognized_namerecognize_faces(frame,self.known_names,self.known_encodings)ifrecognized_name:cv2.putText(frame,recognized_name,(10,30),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,0),2)height,width,channelframe.shape bytes_per_line3*width q_imgQImage(frame.data,width,height,bytes_per_line,QImage.Format_RGB888)self.imageLabel.setPixmap(QPixmap.fromImage(q_img))defstop_camera(self):ifself.cap:self.cap.release()self.timer.stop()defcloseEvent(self,event):self.stop_camera()event.accept()if__name____main__:appQApplication(sys.argv)windowFaceAttendanceApp()window.show()sys.exit(app.exec_())以上就是构建一个基于人工智能的课堂人脸考勤系统的详细步骤和代码示例。你可以根据实际需求调整模型结构、优化超参数或改进界面设计以进一步提升系统的性能和用户体验。
使用OpenCV和Dlib进行人脸特征提取 基于人工智能的课堂人脸考勤系统 使用深度学习模型进行活体检测和专注度检测 使用Qt Designer设计界面基于人工智能的课堂人脸考勤系统文章目录使用OpenCV和Dlib进行人脸特征提取 基于人工智能的课堂人脸考勤系统 使用深度学习模型进行活体检测和专注度检测 使用Qt Designer设计界面1. 环境搭建安装依赖2. 人脸库录入录入人脸库代码3. 人脸识别人脸识别代码4. 活体检测和专注度检测活体检测代码专注度检测代码5. UI设计与结果展示设计界面主程序代码系统主要分为三个部分第一个部分人脸库的录入通过录入人脸裤记录学生信息。第二部分人脸识别选择班级考勤时间后打开摄像头进行人脸考勤。第三部分活体检测和专注度检测判断学生是否在听课以及听课的专注度。1基于人工智能的课堂人脸考勤系统人脸库录入、人脸识别、活体检测和专注度检测。关键代码示例1. 环境搭建确保你的开发环境已经安装了必要的库和工具。安装依赖# 创建并激活虚拟环境conda create-nface_attendancepython3.8conda activate face_attendance# 安装相关库pipinstallopencv-python numpy dlib imutils PyQt5 tensorflow keras scikit-learn2. 人脸库录入使用OpenCV和Dlib进行人脸特征提取并将特征存储在数据库中。录入人脸库代码importcv2importosimportnumpyasnpimportdlibfromimutilsimportface_utilsimportsqlite3# 初始化Dlib的人脸检测器和特征提取器detectordlib.get_frontal_face_detector()predictordlib.shape_predictor(shape_predictor_68_face_landmarks.dat)face_recogdlib.face_recognition_model_v1(dlib_face_recognition_resnet_model_v1.dat)defadd_face_to_db(name,image_path):connsqlite3.connect(faces.db)cconn.cursor()# 创建表如果不存在c.execute(CREATE TABLE IF NOT EXISTS faces (name TEXT, encoding BLOB))# 加载图像并转换为灰度图imagecv2.imread(image_path)graycv2.cvtColor(image,cv2.COLOR_BGR2GRAY)# 检测人脸rectsdetector(gray,0)iflen(rects)0:print(No face detected in the image.)returnforrectinrects:shapepredictor(gray,rect)face_descriptorface_recog.compute_face_descriptor(image,shape)# 将特征向量转换为二进制格式并存储到数据库中face_descriptor_binnp.array(face_descriptor).tobytes()c.execute(INSERT INTO faces (name, encoding) VALUES (?, ?),(name,face_descriptor_bin))conn.commit()conn.close()# 示例添加一张人脸到数据库add_face_to_db(Student1,path/to/student1.jpg)3. 人脸识别使用摄像头实时捕捉图像并与数据库中的人脸特征进行比对。人脸识别代码importcv2importnumpyasnpimportdlibimportsqlite3fromimutilsimportface_utils# 初始化Dlib的人脸检测器和特征提取器detectordlib.get_frontal_face_detector()predictordlib.shape_predictor(shape_predictor_68_face_landmarks.dat)face_recogdlib.face_recognition_model_v1(dlib_face_recognition_resnet_model_v1.dat)defload_faces_from_db():connsqlite3.connect(faces.db)cconn.cursor()c.execute(SELECT name, encoding FROM faces)rowsc.fetchall()known_names[]known_encodings[]forrowinrows:namerow[0]encodingnp.frombuffer(row[1],dtypenp.float64)known_names.append(name)known_encodings.append(encoding)conn.close()returnknown_names,known_encodingsdefrecognize_faces(frame,known_names,known_encodings):graycv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)rectsdetector(gray,0)forrectinrects:shapepredictor(gray,rect)face_descriptorface_recog.compute_face_descriptor(frame,shape)matchesdlib.compare_face_descriptors(known_encodings,face_descriptor,tolerance0.6)ifTrueinmatches:match_indexmatches.index(True)nameknown_names[match_index]print(fRecognized:{name})returnnamereturnNone# 实时人脸识别defreal_time_face_recognition():known_names,known_encodingsload_faces_from_db()capcv2.VideoCapture(0)whileTrue:ret,framecap.read()ifnotret:breakrecognized_namerecognize_faces(frame,known_names,known_encodings)ifrecognized_name:cv2.putText(frame,recognized_name,(10,30),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,0),2)cv2.imshow(Face Recognition,frame)ifcv2.waitKey(1)0xFFord(q):breakcap.release()cv2.destroyAllWindows()# 运行实时人脸识别real_time_face_recognition()4. 活体检测和专注度检测使用深度学习模型进行活体检测和专注度检测。活体检测代码importcv2importnumpyasnpfromkeras.modelsimportload_model# 加载预训练的活体检测模型modelload_model(liveness.model)defis_living_face(frame):face_cascadecv2.CascadeClassifier(haarcascade_frontalface_default.xml)graycv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)facesface_cascade.detectMultiScale(gray,scaleFactor1.1,minNeighbors5,minSize(30,30))iflen(faces)0:(x,y,w,h)faces[0]face_roiframe[y:yh,x:xw]face_roicv2.resize(face_roi,(128,128))face_roiface_roi.astype(float)/255.0face_roinp.expand_dims(face_roi,axis0)predmodel.predict(face_roi)[0]ifpred[1]pred[0]:returnTruereturnFalse# 示例检查当前帧是否为活体人脸capcv2.VideoCapture(0)whileTrue:ret,framecap.read()ifnotret:breakifis_living_face(frame):print(Living face detected!)else:print(Not a living face.)cv2.imshow(Liveness Detection,frame)ifcv2.waitKey(1)0xFFord(q):breakcap.release()cv2.destroyAllWindows()专注度检测代码importcv2importnumpyasnpfromkeras.modelsimportload_model# 加载预训练的专注度检测模型modelload_model(attention.model)defget_attention_level(frame):face_cascadecv2.CascadeClassifier(haarcascade_frontalface_default.xml)graycv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)facesface_cascade.detectMultiScale(gray,scaleFactor1.1,minNeighbors5,minSize(30,30))iflen(faces)0:(x,y,w,h)faces[0]face_roiframe[y:yh,x:xw]face_roicv2.resize(face_roi,(128,128))face_roiface_roi.astype(float)/255.0face_roinp.expand_dims(face_roi,axis0)predmodel.predict(face_roi)[0]attention_levelnp.argmax(pred)returnattention_levelreturnNone# 示例获取当前帧的专注度水平capcv2.VideoCapture(0)whileTrue:ret,framecap.read()ifnotret:breakattention_levelget_attention_level(frame)ifattention_levelisnotNone:print(fAttention Level:{attention_level})cv2.imshow(Attention Detection,frame)ifcv2.waitKey(1)0xFFord(q):breakcap.release()cv2.destroyAllWindows()5. UI设计与结果展示使用PyQt5设计用户界面并展示检测和跟踪结果。设计界面使用Qt Designer设计界面并导出为.ui文件然后使用pyuic5转换为Python代码。pyuic5-xyour_design.ui-oui_design.py主程序代码importsysfromPyQt5.QtWidgetsimportQApplication,QMainWindow,QFileDialogfromPyQt5.QtGuiimportQImage,QPixmapfromPyQt5.QtCoreimportQTimerimportcv2fromui_designimportUi_MainWindowclassFaceAttendanceApp(QMainWindow,Ui_MainWindow):def__init__(self):super().__init__()self.setupUi(self)self.timerQTimer()self.timer.timeout.connect(self.update_frame)self.capNoneself.known_names,self.known_encodingsload_faces_from_db()defstart_camera(self):self.capcv2.VideoCapture(0)self.timer.start(30)defupdate_frame(self):ret,frameself.cap.read()ifret:recognized_namerecognize_faces(frame,self.known_names,self.known_encodings)ifrecognized_name:cv2.putText(frame,recognized_name,(10,30),cv2.FONT_HERSHEY_SIMPLEX,1,(0,255,0),2)height,width,channelframe.shape bytes_per_line3*width q_imgQImage(frame.data,width,height,bytes_per_line,QImage.Format_RGB888)self.imageLabel.setPixmap(QPixmap.fromImage(q_img))defstop_camera(self):ifself.cap:self.cap.release()self.timer.stop()defcloseEvent(self,event):self.stop_camera()event.accept()if__name____main__:appQApplication(sys.argv)windowFaceAttendanceApp()window.show()sys.exit(app.exec_())以上就是构建一个基于人工智能的课堂人脸考勤系统的详细步骤和代码示例。你可以根据实际需求调整模型结构、优化超参数或改进界面设计以进一步提升系统的性能和用户体验。
