This function will classify the student based upon the model which is already trained and stored in database. For prediction, predict() function is called and the classification is done using KNN classifier. Once the prediction is completed, the status of the student is stored in the excel sheet.

image_detection.py

def match():
    destination = str(session.get('destination'))
    print(destination)
    if os.path.isfile(destination):
        test_append = str(session.get('test_append'))
        session['test_append'] = test_append
        id_folder = str(session.get('id_folder'))

        train_dir = APP_ROOT1[0]+"admin_site/train/"+ test_append
        try:
            model = APP_ROOT1[0]+"admin_site/model/"+test_append+"/" + id_folder + "/" +"model"
            print(model)
            return predict1(model)
        except FileNotFoundError:
            os.remove(APP_ROOT1[0]+"teachers_site/image.jpeg")
            return render_template("upload.html",msg="trained model not present for " + test_append + ": "+id_folder)
        

def predict(X_img_path, knn_clf = None, model_save_path ="", DIST_THRESH = .45):
    if knn_clf is None and model_save_path == "":
        raise Exception("must supply knn classifier either thourgh knn_clf or model_save_path")

    if knn_clf is None:
        with open(model_save_path, 'rb') as f:
            knn_clf = pickle.load(f)

    X_img = face_recognition.load_image_file(X_img_path)
    X_faces_loc = face_locations(X_img)
    if len(X_faces_loc) == 0:
        return []

    faces_encodings = face_recognition.face_encodings(X_img, known_face_locations=X_faces_loc)


    closest_distances = knn_clf.kneighbors(faces_encodings, n_neighbors=1)

    is_recognized = [closest_distances[0][i][0] <= DIST_THRESH for i in range(len(X_faces_loc))]
    

    return [(pred) if rec else ("unknown") for pred, rec in zip(knn_clf.predict(faces_encodings), is_recognized)]

def predict1(model):
    test_append = str(session.get('test_append'))
    test_dir = APP_ROOT1[0]+"teachers_site/test/" + test_append
    f_preds = []
    for img_path in listdir(test_dir):
        preds = predict(join(test_dir, img_path) ,model_save_path=model)
        f_preds.append(preds)
        print(f_preds)
    print(len(preds))
    print(len(f_preds))
    for i in range(len(f_preds)):
        if(f_preds[i]==[]):
            os.remove(APP_ROOT1[0]+"teachers_site/image.jpeg")
            return render_template("upload.html",msg="upload again, face not found")
        else:
            os.remove(APP_ROOT1[0]+"teachers_site/image.jpeg")
    if f_preds[0][0] == 'unknown':
        return render_template("upload.html",msg= "Student Not Matched")
    excel = os.path.join(APP_ROOT,"excel/")
    if not os.path.isdir(excel):
        os.mkdir(excel)
    excel1 = os.path.join(excel,test_append)
    if not os.path.isdir(excel1):
        os.mkdir(excel1)
    teacher_name = str(session.get('teacher_name'))
    excel2 = os.path.join(excel1,teacher_name)
    if not os.path.isdir(excel2):
        os.mkdir(excel2)
    session['excel2'] = excel2
    excel3 = excel2+"/"+date+'.xlsx'
    if not os.path.isfile(excel3):
        workbook = xlsxwriter.Workbook(excel2+"/"+date+'.xlsx')
        worksheet = workbook.add_worksheet()
        worksheet.set_column(0,0,20)
        worksheet.write('A1','Roll Id')
        f_preds.sort()
        row = 1
        col = 0
        print(f_preds[0][0], type(f_preds[0][0]))
        for i in range(len(f_preds)):
            for j in range(len(f_preds[i])):
                worksheet.write_string(row,col,f_preds[i][j])
                row += 1
        workbook.close()
        return render_template("upload.html",msg= f_preds[0][0] + " present")
    else:
        df = pd.read_excel(excel2+"/"+date+'.xlsx')
        writer = pd.ExcelWriter(excel2 + "/" + date+'.xlsx')
        df.to_excel(writer,sheet_name="Sheet1",index=False)
        workbook  = writer.book
        worksheet = writer.sheets['Sheet1']
        rows=df.shape[0]
        worksheet.write_string(rows+1,0,f_preds[0][0])
        writer.save()
        df = pd.read_excel(excel2+"/"+date+'.xlsx')
        df.drop_duplicates(['Roll Id'],keep='first',inplace=True)
        # result = df.sort_values("Roll Id")
        writer = pd.ExcelWriter(excel2 + "/" + date+'.xlsx')
        df.to_excel(writer,'Sheet1',index=False)
        workbook = writer.book
        worksheet = writer.sheets['Sheet1']
        worksheet.set_column(0,0,20)
        writer.save()
        return render_template("upload.html",msg= f_preds[0][0] + " present")