Optimization in training required

[up1512001]

I thought optimizing training will be good for our project.

[vinay7102001]

`import cv2
import numpy as np
from keras.models import load_model

import imutils as im

regarding csv file writing

import pandas as pd
import csv
import datetime
import os

def mean_squared_loss(x1, x2): # euclidean distance
difference = x1 - x2
a, b, c, d, e = difference.shape
n_samples = a * b * c * d * e
sq_difference = difference ** 2
Sum = sq_difference.sum()
distance = np.sqrt(Sum)
mean_distance = distance / n_samples # mean squar error
return mean_distance

def help(video):

# csv file handling
file_name = 'E:\\6th_sem\\project\\event detection details.csv'

model = load_model('E:\\6th_sem\\project\\saved_model2.h5')

cap = cv2.VideoCapture(video)
# print(cap.isOpened())
sum = 0
c = 0
while cap.isOpened():
    imagedump = []
    ret, frame = cap.read()
    for i in range(10):
        ret, frame = cap.read()
        if ret == False:
            break
        image = im.resize(frame, width=1000, height=1000)
        frame = cv2.resize(frame, (227, 227), interpolation=cv2.INTER_AREA)
        gray = 0.2989 * frame[:, :, 0] + 0.5870 * frame[:, :, 1] + 0.1140 * frame[:, :, 2]  # convert into gray scale
        gray = (gray - gray.mean()) / gray.std()
        gray = np.clip(gray, 0, 1)
        imagedump.append(gray)
    imagedump = np.array(imagedump)
    imagedump.resize(227, 227, 10)
    imagedump = np.expand_dims(imagedump, axis=0)  # 3d data converted into 4d
    imagedump = np.expand_dims(imagedump, axis=4)  # 4d data into 5d
    output = model.predict(imagedump)
    loss = mean_squared_loss(imagedump, output)
    if ret == False:
        print("video end")
        break
    if cv2.waitKey(10) & 0xFF == ord('q'):
        break

    current_date = datetime.date.today()
    # print(current_date)
    current_time = datetime.datetime.now()
    current_time = current_time.strftime('%H:%M:%S')
    # print(current_time)

    print(loss)

    if loss > 0.00039:
        rows = [[f'{current_date}', f'{current_time}', f'{video[-1:-7:-1][::-1]}', 'Abnormal']]
        # regarding csv file writing
        with open(file_name, 'a') as csvfile:
            csvwriter = csv.writer(csvfile)
            # csvwriter.writerow(fields)
            csvwriter.writerows(rows)
        print('Abnormal Event Detected')
        cv2.putText(image, "abnormal event", (100, 80), cv2.FONT_HERSHEY_TRIPLEX, 2, (250, 24, 255), 4)
    else:
        cv2.putText(image, "normal event", (100, 80), cv2.FONT_HERSHEY_TRIPLEX, 2, (250, 24, 255), 4)
    cv2.imshow("video", image)
cap.release()
cv2.destroyAllWindows()

if name == 'main':
train_video_path = 'E:\6th_sem\project\testing_videos\07.avi'
# train_video_path = 'E:\6th_sem\project\testing_videos\13.avi'
# train_video_path = 'E:\6th_sem\project\testing_videos\12.avi'
# train_video_path = 'E:\6th_sem\project\testing_videos\06.avi'
# train_video_path = 'E:\6th_sem\project\testing_videos\05.avi'
# videos = os.listdir(train_video_path)
# print(videos)
# for video in videos:
# if video.endswith('.mp4') or video.endswith('.avi'):
help(train_video_path)`