detect.py

import cv2
import numpy as np
import pandas as pd
import os
import csv

fname = 'recognizer/trainingData.yml'
if not os.path.isfile(fname):
    print('first train the data')
    exit(0)


names = {}
labels = []
students = []


def from_excel_to_csv():
    df = pd.read_excel('data.xlsx')
    df.to_csv('./data.csv')

def getdata():
    with open('data.csv','r') as f:
        data = csv.reader(f)
        next(data)
        lines = list(data)
        for line in lines:
            names[int(line[0])] = line[1]


def  markPresent(name):
    with open('data.csv','r') as f:
        data = csv.reader(f)
        lines = list(data)
        # for line in lines:
        #     line.pop(0)
        # print(lines)
        for line in lines:
            if line[1] == name:
                line[-1] = 'P'
                with open('data.csv','w') as g:
                    writer = csv.writer(g,lineterminator='\n')
                    writer.writerows(lines)
                    break
    
    # df = pd.read_csv('data.csv')
    # df.to_excel('data.xlsx',index=False)

def update_Excel():
    with open('data.csv') as f:
        data = csv.reader(f)
        lines = list(data)
        for line in lines:
            line.pop(0)
        with open('data.csv','w') as g:
            writer = csv.writer(g,lineterminator='\n')
            writer.writerows(lines)
            
    df = pd.read_csv('data.csv')
    df.to_excel('data.xlsx',index = False)
    

face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
cap = cv2.VideoCapture(0)

# cap.set(3,640) # set Width
# cap.set(4,480) # set Height

from_excel_to_csv() # converting the excel to csv for use
getdata() # getting the data from csv in a dictionary
print('Total students :',names)

recognizer = cv2.face.LBPHFaceRecognizer_create() #LOCAL BINARY PATTERNS HISTOGRAMS Face Recognizer

recognizer.read(fname) # read the trained yml file


while True:   
    ret, img = cap.read()
   # img = cv2.rotate(img, rotateCode=cv2.ROTATE_90_CLOCKWISE)
    #img = cv2.rotate(img, rotateCode=cv2.ROTATE_90_COUNTERCLOCKWISE)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    equ = cv2.equalizeHist(gray) 
    final = cv2.medianBlur(equ, 3)

    faces = face_cascade.detectMultiScale(final, 1.3, 5)
    for (x,y,w,h) in faces:
        cv2.rectangle(img,(x,y),(x+w,y+h),(0,255,0),3)
        label,confidence = recognizer.predict(gray[y:y+h,x:x+w])
        print('label:',label)
        print('confidence:',confidence)
        predicted_name = names[label]
        if confidence < 65:
            confidence = 100 - round(confidence)
            cv2.putText(img, predicted_name +str(confidence) +'%', (x+2,y+h-4), cv2.FONT_HERSHEY_SIMPLEX, 1, (150,255,0),2)
            labels.append(label)
            students.append(names[label])
            totalstudents = set(students)
            justlabels = set(labels)
            print('student Recognised : ',totalstudents,justlabels)
            for i in justlabels:
                if labels.count(i)>20:
                    markPresent(names[label])
   
           

        
        
        
        cv2.imshow('Face Recognizer',img)
        k = cv2.waitKey(30) & 0xff
        if cv2.waitKey(33) == ord('a'):
            print('presed a')
            #update_Excel()
            break



cap.release()
cv2.destroyAllWindows()