Scribd Logo
Upload

Welcome to Scribd!

  • 6 views

    FinalPPT (Do Not Edit)

    Uploaded by

    keerthanktrade1555

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    FinalPPT (Do Not Edit)

    Uploaded by

    keerthanktrade1555
    6 views21 pages

    Original Title

    FinalPPT (Do not edit)

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    6 views21 pages

    FinalPPT (Do Not Edit)

    Uploaded by

    keerthanktrade1555

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 21

    VIDEO-SURVILLANCE-AND-TRACKING-SYSTEM

    USING FACIAL RECOGNITION

    Presented by: Under the guidance of


    Gourav r [4VV20CS042] Dr. RAMAKRISHNA HEGDE
    Hrithik k Gowda [4VV20CS053]
    Associate Professor, Department of Computer Science & Engineering
    Iftheqar Ahmed [4VV20CS054]
    Keerthan K [4VV20CS059]
    ABSTRACT & INTRODUCTION

    CCTV cameras are widely used for monitoring public and private areas, ensuring
    constant surveillance.

    When someone goes missing, CCTV footage becomes essential for search teams.

    It provides clear evidence of where the person was last seen, helping to map out
    their movements and potential routes.

    However, manually going through this footage can be slow and tiresome.

    That's where our project, "The Third Eye," comes in. We're dedicated to simplifying
    this process, making it quicker and real-time.
    SYSTEM ARCHITECTURE

    Server: Backend:
    Apache (Part of Flask) Python
    Web Framework: OpenCV (for image processing)
    Flask (Python) Database:
    Frontend: SQLite3
    Jinja (part of Flask) CSV files
    HTML
    CSS
    Bootstrap 5 (Frontend Framework)
    Software Requirements

    Python 3.11
    Flask Web Framework
    Web Browser

    Hardware Requirements

    Device capable of accessing a web browser (e.g., Laptop,


    Phone)
    CCTV Cameras
    Server with a web server such as Gunicorn, uWSGI, or
    Apache with mod_wsgi
    Work Flow

    Image of the Missing Video stream from CCTV


    Person Cameras

    Save to logs

    Detect the face from the Faces are extracted from


    uploaded picture the Frames

    If Match is
    Found
    Find the encodings of the Store the Data Base with Face Compare the Find the Encodings of the
    face Face Encodings encodings Face Encodings face
    Workflow

    Obtain the CCTV footage, comprising a series of frames, with each frame
    representing an image captured at a defined rate per second.

    Analyze each frame individually to identify any faces present within it.

    Utilize face detection algorithms to accurately locate and extract facial features
    from the detected faces.

    Compare the extracted facial features of each detected face against the stored facial
    data of the missing person within the database.
    Workflow

    Implement facial recognition techniques to determine the likelihood of a match


    between the detected faces and the facial data of the missing person.

    Repeat the process for each frame in the CCTV footage to comprehensively search
    for the missing person across all captured images.

    If the match is found, the timing and location is logged in the database.
    find faces encode Store in
    [1,2,9,5,8,................]
    faces db

    exists in db
    Check if
    find faces encode [8,5,7,9,,6,...............]
    faces [1,2,9,5,8,................]
    face_recognition Library

    The process of extracting Face


    Encodings relies on the
    face_recognition library.

    It is constructed using dlib's cutting-


    edge face recognition technology.

    This model is constructed with deep


    learning techniques and boasts an
    impressive accuracy rate of 99.38%
    on the Labeled Faces in the Wild
    benchmark dataset.
    Dlib

    Dlib is a modern C++ toolkit containing


    machine learning algorithms and tools for
    creating complex software in C++ to solve real
    world problems.
    When a missing persons image is added, the face_recognition library is used to
    encode the image and store it in database

    The Application generates frames from the camera using openv, and face
    recognition has been integrated.

    The face_recognition library is utilized to perform face detection and recognition.

    For each frame captured, faces are detected using the face_recognition function and
    encoded using the face_recognition function.
    The encoded faces are compared with the known faces stored in the database to
    identify any matches

    Once a match is found, a rectangle is drawn around the recognized face, and the
    database is updated accordingly, marking the person as found.
    Labeled Faces in the Wild

    Labeled Faces in the Wild is created and


    maintained by University of Massachusetts.

    Labeled Faces in the Wild is a public


    benchmark for face verification, also known
    as pair matching. No matter what the
    performance of an algorithm on LFW, it
    should not be used to conclude that an
    algorithm is suitable for any commercial
    purpose.
    The Application

    Login Page

    We've set up a simple login page


    using SHA-256 encryption.

    Admins logging in will be directed to


    the admin dashboard.

    Regular users will be directed


    straight to the logs page, limiting
    their access to major actions.
    The Application

    Dashboard

    The admin has access to the dashboard.

    The dashboard features a camera feed.

    Various options are available on the


    dashboard, including adding a new user.

    Another option allows for registering a


    missing person.
    The Application

    Add Users

    The admin has the ability to add a


    new user.

    During the user creation process,


    the admin can designate the user's
    role as either "User" or "Admin".
    The Application

    Manage Users

    The admin can elevate a


    regular user to the admin role.

    Additionally, the admin has


    the authority to remove users
    from the system.
    The Application

    Logs

    The logs contain detailed


    information about the last known
    location of a missing person,
    updating every second to provide
    a brief overview of their
    movements.

    Normal users have access to the


    same dashboard, but without
    access to additional features.
    The Application

    Missing Register

    A clear picture of the missing


    person is included.
    Their name is added for
    identification purposes, facilitating
    their addition to our logs.
    Scope for further improvements

    Utilize Python's multithreading concept to create separate threads for each camera,
    enhancing efficiency and reducing memory consumption.

    Implement validation during the upload of a missing person's picture to ensure that the
    image contains only a single person.

    Verify the picture to confirm whether it contains a single individual before proceeding
    with the upload process.

    Currently, These improvements couldn't be implemented currently


    due to hardware constraints
    Thank You

    You might also like