1. Design and implementation Safety Monitoring
System for Nursery School Children Using Computer
Vision Technology
UNDERGUILDENCE PRESENTED BY
MR.SELVAKUMAR ISHWARYA E - C21UG151CSC078
JANAKI R - C21UG151CSC079
JOTHIGA K - C21UG151CSC080
2. ABSTRACT
• Children’s In nursery schools, accurately analyzing the natural behavior of
active and sensitive children can be challenging when an observer (a person
holding a camera) is present.
• Therefore, we propose the development of a system for analyzing
children’s behavior based on human tracking from live cameras in nursery
schools. The initial phase involves training a model to detect children in
live camera feeds using a Convolutional Neural Network (CNN) algorithm
for person detection.
• To further enhance the system's capabilities, we introduce a feature for
recognizing and capturing abnormal activities, with notifications sent to
authorized personnel through the Internet Message Access Protocol
(IMAP).
• This ensures that the system is not only for passive observation but also for
actively identifying and responding to potentially concerning behaviors in
real-time.
3. System specfication
Hardware Specification
• Processor type : I5 processor
• RAM : 8GB RAM
• Storage : 1TB
• Display : 20’ color display
Software Specification
• Front end : PyQt5
• Back end : python language
• Software tool used : Pycharm
• Platform : Windows 8
4. EXISTING SYSTEM
• In various institutions, the common problem faced is to locate the
staff/student immediately when needed.
• Few methods, which are in use, are the old fashioned traditional
announcement systems.
• The existing system is that the privacy of the staff/student is affected
& it also interrupts the regular functioning of the institution.
Disadvantages:
• Continuous monitoring raises the children and staff, potentially
leading to concerns about unauthorized access and data misuse.
• Implementation and maintenance of a reliable tracking system may
face technical complexities, including reliability issues, accuracy
challenges, and technical glitches.
5. Proposed system
• The proposed system will use multiple methods and integrate computer
vision technology to create a framework for monitoring and analyzing
children's behavior.
• Implementation of person detection algorithms to identify individuals
within the video frames. Integration of object tracking techniques to follow
the movement of each person continuously.
• The real-time notifications to authorized personnel through the Internet
Message Access Protocol (IMAP).
Advantages:
• Utilizes live camera feeds and computer vision to continuously monitor
children's activities in nursery schools. Enhances safety by promptly
identifying and responding to abnormal behaviors.
• Incorporates abnormal activity recognition to enable proactive surveillance.
Sends real-time notifications to authorized personnel via IMAP for swift
intervention in potential safety concerns.
6. Architecture design
welcome page Login page
Monitoring page
Normal
Camera
Segmentation process
preprocessing
Monitor Activities
CNN Algorithm
Student detect
Abnormal
Continue
monitoring Capture image
IMAP Protocol
Authorised person
DBMS
7. Modules
In this project there are 5 module
• Welcome module
• Input camera module
• Person detection module
• Database module
• Recognition Module
• Output module
8. Modules Description
• Welcome module:
The welcome module is the start on click to accessing the next page that is login
page, 2 text boxes for user name and password only username and password are
correct, we can go to next page on clicking log in. If the entered username or
password is incorrect, indicating "Wrong username or password”.
• Input camera module
After successfully login on clicking monitor the camera start capturing 25 fps in
each frame are then preprocessed and analyzed of image from video processing.
• Person detection module:
The Person Detection Module is a vital component of the Nursery School
Behavioral Analysis System, utilizing advanced computer vision algorithms for
real-time identification of individuals in nursery school environments. Performance
metrics are monitored to ensure efficient operation, making it a foundational
element for understanding and monitoring the behavior of nursery school children.
9. Cont..,
• Database module:
A database is an organized collection of image and train dataset information, or
data, spatial data for subsequent tracking and analysis stored computer system. A
database is usually controlled by a database management system (DBMS).
• Recognition Module:
The Recognition Module in the Nursery School Behavioral Analysis System
employs and predefined behavior patterns to identify abnormal activities in real-
time. It collaborates with output of trained CNN model for Person Detection and
Tracking modules, if any abnormal behavior is detected image is captured
immediately.
• Output Module:
The captured image along with the notifications is send through IMAP to
authorized personnel for immediate actions. This Output Module empowers
authorized personnel with actionable information, fostering informed decision-
making and proactive surveillance for the safety and well-being of nursery school
children.
11. RESULT
The implementation and rigorous testing of our nursery school
behavior analysis system, the results demonstrate a robust and effective
solution for enhancing the safety and well-being of children in nursery
school environments. The person detection module accurately identifies
and locates children within live camera feeds, providing a reliable
foundation for subsequent tracking.
The tracking module ensures continuous monitoring of
children's positions in real-time, while the abnormal activity
recognition module effectively analyzes behavior pa The integration of
real-time notifications through the Simple Mail Transfer Protocol
(SMTP) further enhances the system's capabilities, enabling prompt
alerts to authorized personnel in the event of abnormal activities.