Finding Missing Persons with AI

1.
INTRODUCTION
1.1 .INTRODUCTION
Face recognition is a biometric-based technology

that mathematically maps a particular person’s or individual’s
facial features and stores all that data as a face print.By
usingthis technique, the information of the face of a person is
savedmathematically or in the format of graphs in the
database,which is used for detecting that particular
face.Facerecognition model in our system will find a match of
that person in the database. If a match is found, it
will be notified tothe police and the guardian of that person.

In this paper we will use the ideas of the AWS facialrecognition
algorithm which is based on Artificial Intelligence(AI) and will detect
faces with the maximum accuracies to findthe missing person.
In the world, a countless number of people are missing

everyday which includes kids, teens, mentally challenged, old-
agedpeople with Alzheimer's, etc. Most of them remain
untraced.This paper proposes a system that would help the
police andthe public by accelerating the process of searching
using facerecognition.Face recognition technique can be used
for many things andfinding the missing person is a biggest
advantage for any facerecognition technique. To make the task
of finding themissing person easier we are planning to make
anapplication which will be accessed by some
volunteersthrough which we can find missing person in short
span oftime. This will make the work of police to find a
particularperson easier.Meanwhile, there is a need of
automation for automating thetask of finding the particular
person by recognizingparticular image and comparing that
image with other imagein order to check whether both images
has samecharacteristics or not. By doing this we will come to
knowwhether the missing person in the image clicked
fromparticular location is correct or not, and if it is correct
thenpolice can start their next steps to find the person from
thatarea.
1.2 . Project Overview:
The aim of this project is to develop a system that uses

AI to help locate missing persons. The system will take
input from various sources, such as social media, news
reports, and law enforcement databases, and use
natural language processing, image recognition, and
machine learning algorithms to analyze the data and
identify potential leads on the whereabouts of missing
persons. The project will involve collecting and cleaning
large amounts of data, developing and training AI
models, and creating a user interface to present the
results. Develop a system that can analyze data from
various sources to identify potential leads on the
whereabouts of missing persons
Use natural language processing to analyze social
media posts and news reports related to missing
persons
Use image recognition to analyze photos and videos
related to missing persons
Use machine learning algorithms to identify patterns
and connections in the data
Create a user interface to present the results to users
When we went through the website, we immediately
understood the issue. The process to submit pictures of
a
child (you find suspicious) in your area is tricky and not
anonymous.
People who employ these children are powerful people
nobody wants to mess with; this is why the user prefers
anonymous submission.
The initiative wasn’t using the power of machine
learning.
Since it is happening on a large scale, there should be
an
automated solution.
As shown in below image we can access the all
information
of missing person under the tab of ‘Photographs of
Missing
persons’ as well as we can access the photographs of
recovered children under the tab of ‘Photographs of
Recovered children.
1.3. Methodologies
This is the high-level system Design of our proof of concept which uses AWS Rekognition for facial
recognition. The picture of a missing person will be uploaded to the application with the help of AWS
Rekognition. AWS Rekoginition will identify a face in that picture and then that face ID along with the
missing details will be stored in the S3 bucket. Then the search of the user's human face will happen
by uploading an image. If this new image has a face already available within the database, then AWS
recognition will trigger an SNS notification to an email id. That's the high-level design, the first
business process by using volunteers to capture an image of a human being. Then we will store that
image in the application with the help of AWS Rekognition along with some other details of the
human being. Those details will be stored in the S3 bucket. AWS Rekognition will analyze the image
uploaded. It will return a unique face id which is always unique for a given face and then that face id
will be used as a unique key to store the related details of the missing person in DynamoDB and once
we want to search the face, we will use an existing Image. Using the search functionality, that new
face image will be uploaded to the application with AWS recognition. If a face match is found, an SNS
notification will be triggered and an email will be sent to the respective authority’s email id
databases
Clean and preprocess the data to remove noise and
irrelevant information
Use natural language processing to analyze text data
and identify keywords and patterns related to missing
persons
related to missing persons and identify potential
matches with missing persons databases
Use machine learning algorithms, such as clustering
and classification, to identify patterns and connections
in the data
Develop a user interface to present the results to users
2. Technologies and Techniques
2.1 . Natural Language Processing (NLP):

NLP can be used to analyze text data from
various sources, such as social media and news
reports, to identify relevant information related
to missing persons. NLP techniques like
sentiment analysis, topic modeling, and named
entity recognition can help in identifying
keywords and patterns related to missing
persons
2.2 Image Recognition:

Image recognition can be used to analyze photos
and videos related to missing persons and match
them with missing persons databases. Techniques
like object detection, face recognition, and image
classification can be used to identify relevant
images and videos related to missing persons.
Machine Learning: Machine learning algorithms

like clustering, classification, and regression can
be used to identify patterns and connections in
the data related to missing persons. These
algorithms can help in predicting the location or
behavior of missing persons based on historical
data.
2.3 .Data Mining:

Data mining techniques can be used to collect
and preprocess data from various sources, such
as social media, news reports, and law
enforcement databases. These techniques can
help in cleaning and transforming data into a
usable format for analysis.
Graph Databases: Graph databases can be used

to store and analyze the relationships between
entities related to missing persons, such as
people, places, and events. Graph databases can
help in identifying connections and patterns in
the data that are not easily identifiable using
traditional databases.
User Interface Design: A user interface can be

designed to present the results of the analysis to
users. The user interface should be designed in
such a way that it is easy to use and understand
for users who are not familiar with AI or data
analysis.
Cloud Computing: Cloud computing can be used

to store and process large amounts of data
related to missing persons. Cloud computing
platforms like Amazon Web Services (AWS) and
Microsoft Azure provide scalable and secure
infrastructure for data storage and analysis.
3. Requirement and Analysis
3.1 Problem Statement:

Develop a system that can analyze data from
persons
Use image recognition to analyze photos and
videos related to missing persons
Use machine learning algorithms to identify
patterns and connections in the data
Create a user interface to present the results to
users
Manual System for finding missing person have
very
long procedure and takes more time. More time is
require for launching an FIR (First Information
Report) in police station. Also time required for
finding lost person is more. Also during manual
process number of manpower for searching lost
person
is less. Some existing application does not show
the
proper information about the Missing person,
which is
difficult to find out missing person [9]. Some
missing
person related website only shows the database of
missing person.In some amusement parks and
other
public venues, they have used RFID chip for
searching
Missing Children. [2]
To overcome from this, there exist some android
applications has been developed. But these
applications has certain limitations as
3.2 Scope of the System:

To overcome from these drawbacks we are developing
Proposed System as "Finding Missing Person (FMP)".
This application is basically designed to perform all
the tasks that previous system can perform all
functionalities that are provided by existing
applications as well as it gives additional feature to
user. It will be for all android devices which support at
least Android 2.1 Platform. We got idea about how
interface should be for adding new complaint (How
add complaint form user) from this android
application.
Proposed System will contain following features:
 Display Information about missing person.
 Adding new complaint.
 Removing Complaints.
 Searching person by particular attribute such
as name, location etc.
 Notification Portal.
The proposed system contains the following Modules:

Sign In/Sign Up Activity:
 User will first go to sign in fragment if he/she has
not created profile then user will go to sign up.
 In Sign Up user will have to enter username, email
and password.
 After entering this user will receive verification link
on email and user will have to click on that link to
get verified.
 After authentication user’s profile will get created.
 Police also sign up using same method just they
need to enter their location(Googlemap integrated)
with mobile number so that their profile will get
created to that specific location on Google map.
 User can sign in into the account.
Add Report/Case Activity:
 Here anybody will be able report the missing
person.
 User need to enter missing persons details like
name, age, height etc. with the location
 User can select exact location with Google map
integration.
 Also need to upload image of missing person for
face detection.
 This will create missing persons profile and it will
get added in missing persons list.
Detect Face Activity:
 In this activity user will be able to match the faces.
 User need to hold the camera in front of suspicious
person who he thinks that is missing.
 If the match found in cloud database that is firebase
then that user will be redirected to profile of that
missing person.
 On profile there is location of that person with
reporter’s mobile number and other details
3.3 Functional Requirements:

Collect data from various sources, such as social media,
news reports, and law enforcement databases
persons
in the data
3.4 Diliverables:
A system that can analyze data from various sources to
persons
Image recognition with the use of one-shot learning
has become very powerful. This technology when put
into
good use, can be beneficial. It can even be used in
Hotels,
Hospitals, etc., to find criminals instantly.
Process of identifying the missing people is fastened.
Our
system replaces the manual scanning process through
the
databases for each picture to check the match, by an
efficient
face recognition method which finishes the work in no
time.
It will be useful to get exact location of the person if
match
detected with the Google maps integration which also
makes
police job easy. it will be helpful to contact police
quickly as
well.
By using the TensorFlow Face recognition we are trying
to
achieve almost 77.99% accuracy with the help of pre-
trained
model.
In the future, there is a scope to extend this system
further
by connecting our system to public cameras and detect
faces
real-time. The frames will be continuously sent by the
public
cameras to our system where our system will be
continually
monitoring the frames. When a lost person is identified
in
any of the frames, It will notify the concerned
authorities, the
method that finishes the work in no time.
A user interface to present the results to users
4. System Design:
4.1. System Design:

application
1) Presentation Layer:
It is front end component, which is responsible for
providing portable presentation logic. Mobile phone
will act as thin client. Phone will contain Application.
User will interact with application to add complaint
and send this data to web service
2) Business Layer (Web Service):
The business layer function (web service) between
presentation layer and Database layer sending the
client’s request to database. Web service will be
responsible to fetch data from client, process it and
then store it in database. Web service act as
middleware for Application and Database. In our
project used JSON web service for connectivity.
[5][11]
3) Database Layer:
Database is responsible for storing all information in
well-defined format. Also it responses to the queries
fired by client to add, update, remove or search
records. In Our project we have used PHP, MYSQL
database for storing Information.[3]
4.2. UseCase Diagram:

In this use case diagram, the User Interface allows
users to search for missing persons. The Data Collection
& Preprocessing System collects and preprocesses data
from various sources, such as social media, news
reports, and law enforcement databases. The AI
Analysis System uses natural language processing,
image recognition, and machine learning algorithms to
analyze the data and identify potential leads on the
whereabouts of missing persons. The Results System
displays the results of the analysis to the user in a user-
friendly format.
Note that this is just an example use case diagram, and

the specific use cases and components of the system
may vary depending on the requirements of the
project.
4.3. Class Diagram:

In this class diagram, the User class represents the user
interface, which has methods for searching for missing
persons and viewing results. The Data Collection &
Preprocessing class is responsible for collecting and
preprocessing data from various sources. The AI
Analysis System class is responsible for analyzing the
data using an algorithm, and the Results System class is
responsible for displaying the results to the user.
Finally, the Algorithm class represents the various AI

algorithms that can be used for analyzing the data.
Note that this is just an example class diagram, and the
specific classes and methods may vary depending on
the requirements of the project.
4.4. Sequence Diagram:
In this sequence diagram, the User Interface initiates a
search for missing persons by sending a request to the
Data Collection & Preprocessing System. The system
then collects data from various sources and
preprocesses it before sending it to the AI Analysis
System. The AI Analysis System analyzes the data using
algorithms such as natural language processing and
image recognition, and returns the results to the
Results System. The Results System then displays the
results to the user.
4.5. Activity Diagram:

In this activity diagram, the process starts with the user
searching for missing persons. The system then collects
data from various sources, such as law enforcement
databases and social media, and preprocesses it to
prepare it for analysis. The preprocessed information is
then analyzed using AI algorithms, such as natural
language processing and image recognition, to identify
potential leads. Finally, the results are displayed to the
user, which could include information on potential
locations, sightings, or other relevant data.
4.6. Deployment Diagram:

In this deployment diagram, the User Interface
Application is deployed on one or more servers and
communicates with the Data Collection &
Preprocessing System to collect data from various
sources. The Data Collection & Preprocessing System is
responsible for preprocessing the data before sending
it to the AI Analysis System. The AI Analysis System is
responsible for analyzing the data using AI algorithms
such as natural language processing and image
recognition.
The Results System is responsible for displaying the

results to the user and is deployed on one or more
servers. The Load Balancer is responsible for
distributing the incoming user requests to the
appropriate server.
5. Implementation:
5.1 Implementaion :
The proposed system, which we have designed,works
on below methodology.Here we are detecting the facial
key points of lost person;dlib facial landmark produces
about 68 unique points for a face. It uses a one-shot
learning technique. Those produced points are float
values with accuracy of about 8 points afterthe
decimal. It is something like this
When the desktop application registers about

threecases, dlib generates 136 * 3 facial landmark
points. 136 is the count as x, y coordinate would be
generated for each point, there are 68 points,and
threeis the number of cases.Then the classifier is
trained based on these points. Suppose if KNN
dispense the facial landmark points in the plot
where RED corresponds to Person 1, GREEN for
Person 2 and BLUE for Person
After this for matching any person’s face,we need to
simply produce facial landmarks of that person and we
will then try to predict the confidence using the KNN
classifier, whichwe trained, if the confidence goes
above 60 then it issaid that it is the same person.Now,
if a person is matched using the KNN model,we trained
then as output,we will get the encoded label. This
encoded label will contain information of that perso
5.2. Modules:
Possible modules for a project that uses AI to find
missing persons:
1. Data Collection Module: This module is responsible
for collecting data from various sources, such as
social media, news articles, and law enforcement
agencies. The data could include information
about missing persons, last known locations, and
potential leads.
2. Data Preprocessing Module: This module is
responsible for cleaning and preparing the data for
analysis. This could involve tasks such as data
cleaning, feature extraction, and data
transformation.
3. Natural Language Processing (NLP) Module: This
module is responsible for analyzing text data, such
as social media posts and news articles, to extract
relevant information about missing persons. This
could include techniques such as named entity
recognition and sentiment analysis.
4. Image and Video Analysis Module: This module is
responsible for analyzing images and video footage
to identify missing persons or potential leads. This
could include techniques such as object detection,
face recognition, and facial expression analysis.
5. Machine Learning Module: This module is
responsible for building and training machine
learning models to identify patterns and potential
leads in the data. This could include techniques
such as clustering, classification, and regression.
6. Geospatial Analysis Module: This module is
responsible for analyzing location data, such as
GPS data or crime incident reports, to identify
potential areas where missing persons may be
located. This could include techniques such as heat
mapping and spatial clustering.
7. Decision Support System Module: This module is
responsible for providing decision support to law
enforcement agencies and other stakeholders
based on the results of the analysis. This could
include generating reports, visualizations, and
recommendations based on the data.
8. User Interface Module: This module is responsible
for providing a user-friendly interface for law
enforcement agencies and other stakeholders to
access the results of the analysis and make
informed decisions.
5.3. Algorithms:
The k-Nearest Neighbors (kNN) algorithm is a machine
learning algorithm that could potentially be used in a
project that uses AI to find missing persons.
The kNN algorithm is a type of supervised learning
algorithm that can be used for classification or
regression. In the case of finding missing persons, it
could be used for classification, where the algorithm is
trained on a dataset of known missing persons and
non-missing persons to make predictions about new
cases.
Here's how the kNN algorithm works:
The algorithm is trained on a labeled dataset of missing

persons and non-missing persons. Each data point in
the dataset includes features such as age, gender, last
known location, and other relevant information.
When a new case of a missing person is reported, the

algorithm analyzes the features of the case and
calculates the distance between the new case and the
existing cases in the training dataset.
The kNN algorithm then selects the k nearest cases

based on their distance to the new case, where k is a
user-defined parameter.
The algorithm classifies the new case based on the
majority label of the k nearest cases. For example, if
the majority of the k nearest cases are labeled as
missing persons, the algorithm would classify the new
case as a missing person.
The kNN algorithm is relatively simple and easy to

implement, and it can be effective for classification
tasks with small to medium-sized datasets. However, it
may not perform as well with larger datasets or when
the features are not well-defined or informative.
Note that the kNN algorithm is just one of many

machine learning algorithms that could potentially be
used in a project that uses AI to find missing persons.
The choice of algorithm would depend on the specific
requirements of the project and the type of data that is
available.
6.Test Cases
Test Cases for finding missing person using AI:
1. Test case: Using facial recognition technology
Input: A set of images of the missing person and a

database of images
Output: A match of the missing person's image with
one of the images in the database
2. Test case: Using natural language processing
Input: A description of the missing person's physical

appearance, last known location, and any other
relevant details
Output: A list of potential matches based on the
description
3. Test case: Using location tracking
Input: GPS data from the missing person's last known

location and any subsequent locations, as well as a
database of locations
location data
4. Test case: Using social media monitoring
Input: Social media posts, messages, and other data

related to the missing person, as well as a database of
social media activity
Output: A list of potential matches based on the social
media data
5.Test case: Using machine learning algorithms
Input: A variety of data sources, including images,

location data, social media data, and any other relevant
information
Output: A list of potential matches based on patterns
and correlations in the data
1.Using facial recognition technology: This test case
involves using a facial recognition algorithm to
compare the missing person's image with a database of
images. The algorithm would analyze the features of
the person's face and compare them to other images in
the database to identify potential matches. This test
case requires a high-quality image of the missing
person and a database of images that includes
potential matches.
2. Using natural language processing: This test case

involves using natural language processing algorithms
to analyze a description of the missing person's
physical appearance, last known location, and any
other relevant details. The algorithm would analyze the
description and compare it to a database of
descriptions to identify potential matches. This test
case requires a detailed and accurate description of the
missing person and a database of descriptions that
includes potential matches.
3.Using location tracking: This test case involves using

location data from the missing person's last known
location and any subsequent locations to identify
potential matches. The algorithm would analyze the
location data and compare it to a database of locations
to identify potential matches. This test case requires
accurate location data and a database of locations that
4.Using social media monitoring: This test case involves

using social media data related to the missing person,
including posts, messages, and other data, to identify
social media data and compare it to a database of
social media activity to identify potential matches. This
test case requires access to the missing person's social
media accounts or public data related to the missing
person, as well as a database of social media activity
that includes potential matches.
5.Using machine learning algorithms: This test case

involves using machine learning algorithms to analyze a
variety of data sources, including images, location data,
social media data, and any other relevant information,
to identify potential matches. The algorithm would
analyze patterns and correlations in the data to identify
potential matches. This test case requires a diverse set
of data sources and a database of potential matches
that includes a variety of data types. It also requires
training the machine learning algorithm with labeled
data to improve accuracy.
6.Using facial recognition technology:
Input: A high-quality image of the missing person and a

database of images.
Algorithm: The algorithm would use a neural network
to analyze the features of the missing person's face and
compare them to other images in the database to
identify potential matches. It would use deep learning
techniques to learn the patterns and features that
distinguish different faces.
Output: The output would be a match of the missing
person's image with one of the images in the database.
7.Using natural language processing:
Input: A detailed and accurate description of the

missing person's physical appearance, last known
location, and any other relevant details, as well as a
database of descriptions.
Algorithm: The algorithm would use natural language
processing techniques to analyze the description and
identify key features, such as height, weight, hair color,
and clothing. It would compare these features to other
descriptions in the database to identify potential
matches.
Output: The output would be a list of potential
matches based on the description.
8.Using location tracking:

database of locations.
Algorithm: The algorithm would use machine learning
techniques to analyze the location data and identify
patterns and correlations. It would compare the
location data to other locations in the database to
identify potential matches.
matches based on the location data.
9.Using social media monitoring:

social media activity.
processing techniques to analyze the social media data
and identify key features, such as interests, activities,
and connections. It would compare these features to
other social media data in the database to identify
potential matches.
matches based on the social media data.
10.Using machine learning algorithms:

information, as well as a database of potential
matches.
Algorithm: The algorithm would use a combination of
machine learning techniques, such as neural networks,
decision trees, and clustering, to analyze the data and
identify patterns and correlations. It would use
supervised learning to train the algorithm with labeled
data and unsupervised learning to identify patterns in
unlabeled data.
matches based on the data sources and the patterns
and correlations identified by the algorithm.
7. Results
8. Conclusion
5. INTRODUCTION
5.1 .INTRODUCTION
Face recognition is a biometric-based technology

that mathematically maps a particular person’s or individual’s
facial features and stores all that data as a face print.By
usingthis technique, the information of the face of a person is
savedmathematically or in the format of graphs in the
database,which is used for detecting that particular
face.Facerecognition model in our system will find a match of
that person in the database. If a match is found, it
will be notified tothe police and the guardian of that person.

In this paper we will use the ideas of the AWS facialrecognition
algorithm which is based on Artificial Intelligence(AI) and will detect
faces with the maximum accuracies to findthe missing person.
In the world, a countless number of people are missing

everyday which includes kids, teens, mentally challenged, old-
agedpeople with Alzheimer's, etc. Most of them remain
untraced.This paper proposes a system that would help the
police andthe public by accelerating the process of searching
using facerecognition.Face recognition technique can be used
for many things andfinding the missing person is a biggest
advantage for any facerecognition technique. To make the task
of finding themissing person easier we are planning to make
anapplication which will be accessed by some
volunteersthrough which we can find missing person in short
span oftime. This will make the work of police to find a
particularperson easier.Meanwhile, there is a need of
automation for automating thetask of finding the particular
person by recognizingparticular image and comparing that
image with other imagein order to check whether both images
has samecharacteristics or not. By doing this we will come to
knowwhether the missing person in the image clicked
fromparticular location is correct or not, and if it is correct
thenpolice can start their next steps to find the person from
thatarea.
1.2 . Project Overview:
The aim of this project is to develop a system that uses

AI to help locate missing persons. The system will take
input from various sources, such as social media, news
reports, and law enforcement databases, and use
natural language processing, image recognition, and
machine learning algorithms to analyze the data and
persons. The project will involve collecting and cleaning
large amounts of data, developing and training AI
models, and creating a user interface to present the
results. Develop a system that can analyze data from
persons
related to missing persons
Use machine learning algorithms to identify patterns
and connections in the data
Create a user interface to present the results to users
When we went through the website, we immediately
understood the issue. The process to submit pictures of
a
child (you find suspicious) in your area is tricky and not
anonymous.
People who employ these children are powerful people
nobody wants to mess with; this is why the user prefers
anonymous submission.
The initiative wasn’t using the power of machine
learning.
Since it is happening on a large scale, there should be
an
automated solution.
As shown in below image we can access the all
information
of missing person under the tab of ‘Photographs of
Missing
persons’ as well as we can access the photographs of
recovered children under the tab of ‘Photographs of
Recovered children.
1.3. Methodologies
This is the high-level system Design of our proof of concept which uses AWS Rekognition for facial
recognition. The picture of a missing person will be uploaded to the application with the help of AWS
Rekognition. AWS Rekoginition will identify a face in that picture and then that face ID along with the
missing details will be stored in the S3 bucket. Then the search of the user's human face will happen
by uploading an image. If this new image has a face already available within the database, then AWS
recognition will trigger an SNS notification to an email id. That's the high-level design, the first
business process by using volunteers to capture an image of a human being. Then we will store that
image in the application with the help of AWS Rekognition along with some other details of the
human being. Those details will be stored in the S3 bucket. AWS Rekognition will analyze the image
uploaded. It will return a unique face id which is always unique for a given face and then that face id
will be used as a unique key to store the related details of the missing person in DynamoDB and once
we want to search the face, we will use an existing Image. Using the search functionality, that new
face image will be uploaded to the application with AWS recognition. If a face match is found, an SNS
notification will be triggered and an email will be sent to the respective authority’s email id
databases
persons
in the data
6. Technologies and Techniques
6.1 . Natural Language Processing (NLP):

NLP can be used to analyze text data from
various sources, such as social media and news
reports, to identify relevant information related
to missing persons. NLP techniques like
sentiment analysis, topic modeling, and named
entity recognition can help in identifying
keywords and patterns related to missing
persons
6.2 Image Recognition:

Image recognition can be used to analyze photos
and videos related to missing persons and match
them with missing persons databases. Techniques
like object detection, face recognition, and image
classification can be used to identify relevant
images and videos related to missing persons.
Machine Learning: Machine learning algorithms

like clustering, classification, and regression can
be used to identify patterns and connections in
the data related to missing persons. These
algorithms can help in predicting the location or
behavior of missing persons based on historical
data.
2.3 .Data Mining:

Data mining techniques can be used to collect
and preprocess data from various sources, such
as social media, news reports, and law
enforcement databases. These techniques can
help in cleaning and transforming data into a
usable format for analysis.
Graph Databases: Graph databases can be used

to store and analyze the relationships between
entities related to missing persons, such as
people, places, and events. Graph databases can
help in identifying connections and patterns in
the data that are not easily identifiable using
traditional databases.
User Interface Design: A user interface can be

designed to present the results of the analysis to
users. The user interface should be designed in
such a way that it is easy to use and understand
for users who are not familiar with AI or data
analysis.
Cloud Computing: Cloud computing can be used

to store and process large amounts of data
related to missing persons. Cloud computing
platforms like Amazon Web Services (AWS) and
Microsoft Azure provide scalable and secure
infrastructure for data storage and analysis.
7. Requirement and Analysis
7.1 Problem Statement:

Develop a system that can analyze data from
persons
Use image recognition to analyze photos and
videos related to missing persons
Use machine learning algorithms to identify
patterns and connections in the data
Create a user interface to present the results to
users
Manual System for finding missing person have
very
long procedure and takes more time. More time is
require for launching an FIR (First Information
Report) in police station. Also time required for
finding lost person is more. Also during manual
process number of manpower for searching lost
person
is less. Some existing application does not show
the
proper information about the Missing person,
which is
difficult to find out missing person [9]. Some
missing
person related website only shows the database of
missing person.In some amusement parks and
other
public venues, they have used RFID chip for
searching
Missing Children. [2]
To overcome from this, there exist some android
applications has been developed. But these
applications has certain limitations as
7.2 Scope of the System:

application.
Proposed System will contain following features:
 Display Information about missing person.
 Adding new complaint.
 Removing Complaints.
 Searching person by particular attribute such
as name, location etc.
 Notification Portal.
The proposed system contains the following Modules:

Sign In/Sign Up Activity:
 User will first go to sign in fragment if he/she has
not created profile then user will go to sign up.
 In Sign Up user will have to enter username, email
and password.
 After entering this user will receive verification link
on email and user will have to click on that link to
get verified.
 After authentication user’s profile will get created.
 Police also sign up using same method just they
need to enter their location(Googlemap integrated)
with mobile number so that their profile will get
created to that specific location on Google map.
 User can sign in into the account.
Add Report/Case Activity:
 Here anybody will be able report the missing
person.
 User need to enter missing persons details like
name, age, height etc. with the location
 User can select exact location with Google map
integration.
 Also need to upload image of missing person for
face detection.
 This will create missing persons profile and it will
get added in missing persons list.
Detect Face Activity:
 In this activity user will be able to match the faces.
 User need to hold the camera in front of suspicious
person who he thinks that is missing.
 If the match found in cloud database that is firebase
then that user will be redirected to profile of that
missing person.
 On profile there is location of that person with
reporter’s mobile number and other details
7.3 Functional Requirements:

Collect data from various sources, such as social media,
news reports, and law enforcement databases
persons
in the data
7.4 Diliverables:
A system that can analyze data from various sources to
persons
Image recognition with the use of one-shot learning
has become very powerful. This technology when put
into
good use, can be beneficial. It can even be used in
Hotels,
Hospitals, etc., to find criminals instantly.
Process of identifying the missing people is fastened.
Our
system replaces the manual scanning process through
the
databases for each picture to check the match, by an
efficient
face recognition method which finishes the work in no
time.
It will be useful to get exact location of the person if
match
detected with the Google maps integration which also
makes
police job easy. it will be helpful to contact police
quickly as
well.
By using the TensorFlow Face recognition we are trying
to
achieve almost 77.99% accuracy with the help of pre-
trained
model.
In the future, there is a scope to extend this system
further
by connecting our system to public cameras and detect
faces
real-time. The frames will be continuously sent by the
public
cameras to our system where our system will be
continually
monitoring the frames. When a lost person is identified
in
any of the frames, It will notify the concerned
authorities, the
method that finishes the work in no time.
A user interface to present the results to users
8. System Design:
4.7. System Design:

application
1) Presentation Layer:
It is front end component, which is responsible for
providing portable presentation logic. Mobile phone
will act as thin client. Phone will contain Application.
User will interact with application to add complaint
and send this data to web service
2) Business Layer (Web Service):
The business layer function (web service) between
presentation layer and Database layer sending the
client’s request to database. Web service will be
responsible to fetch data from client, process it and
then store it in database. Web service act as
middleware for Application and Database. In our
project used JSON web service for connectivity.
[5][11]
3) Database Layer:
Database is responsible for storing all information in
well-defined format. Also it responses to the queries
fired by client to add, update, remove or search
records. In Our project we have used PHP, MYSQL
database for storing Information.[3]
4.8. UseCase Diagram:

In this use case diagram, the User Interface allows
users to search for missing persons. The Data Collection
& Preprocessing System collects and preprocesses data
from various sources, such as social media, news
reports, and law enforcement databases. The AI
Analysis System uses natural language processing,
image recognition, and machine learning algorithms to
analyze the data and identify potential leads on the
whereabouts of missing persons. The Results System
displays the results of the analysis to the user in a user-
friendly format.
Note that this is just an example use case diagram, and

the specific use cases and components of the system
may vary depending on the requirements of the
project.
4.9. Class Diagram:

In this class diagram, the User class represents the user
interface, which has methods for searching for missing
persons and viewing results. The Data Collection &
Preprocessing class is responsible for collecting and
preprocessing data from various sources. The AI
Analysis System class is responsible for analyzing the
data using an algorithm, and the Results System class is
responsible for displaying the results to the user.
Finally, the Algorithm class represents the various AI

algorithms that can be used for analyzing the data.
Note that this is just an example class diagram, and the
specific classes and methods may vary depending on
the requirements of the project.
4.10. Sequence Diagram:
In this sequence diagram, the User Interface initiates a
search for missing persons by sending a request to the
Data Collection & Preprocessing System. The system
then collects data from various sources and
preprocesses it before sending it to the AI Analysis
System. The AI Analysis System analyzes the data using
algorithms such as natural language processing and
image recognition, and returns the results to the
Results System. The Results System then displays the
results to the user.
4.11. Activity Diagram:

In this activity diagram, the process starts with the user
searching for missing persons. The system then collects
data from various sources, such as law enforcement
databases and social media, and preprocesses it to
prepare it for analysis. The preprocessed information is
then analyzed using AI algorithms, such as natural
language processing and image recognition, to identify
potential leads. Finally, the results are displayed to the
user, which could include information on potential
locations, sightings, or other relevant data.
4.12. Deployment Diagram:

In this deployment diagram, the User Interface
Application is deployed on one or more servers and
communicates with the Data Collection &
Preprocessing System to collect data from various
sources. The Data Collection & Preprocessing System is
responsible for preprocessing the data before sending
it to the AI Analysis System. The AI Analysis System is
responsible for analyzing the data using AI algorithms
such as natural language processing and image
recognition.
The Results System is responsible for displaying the

results to the user and is deployed on one or more
servers. The Load Balancer is responsible for
distributing the incoming user requests to the
appropriate server.
5. Implementation:
5.1 Implementaion :
The proposed system, which we have designed,works
on below methodology.Here we are detecting the facial
key points of lost person;dlib facial landmark produces
about 68 unique points for a face. It uses a one-shot
learning technique. Those produced points are float
values with accuracy of about 8 points afterthe
decimal. It is something like this
When the desktop application registers about

threecases, dlib generates 136 * 3 facial landmark
points. 136 is the count as x, y coordinate would be
generated for each point, there are 68 points,and
threeis the number of cases.Then the classifier is
trained based on these points. Suppose if KNN
dispense the facial landmark points in the plot
where RED corresponds to Person 1, GREEN for
Person 2 and BLUE for Person
After this for matching any person’s face,we need to
simply produce facial landmarks of that person and we
will then try to predict the confidence using the KNN
classifier, whichwe trained, if the confidence goes
above 60 then it issaid that it is the same person.Now,
if a person is matched using the KNN model,we trained
then as output,we will get the encoded label. This
encoded label will contain information of that perso
5.2. Modules:
Possible modules for a project that uses AI to find
missing persons:
9. Data Collection Module: This module is responsible
for collecting data from various sources, such as
social media, news articles, and law enforcement
agencies. The data could include information
about missing persons, last known locations, and
potential leads.
10. Data Preprocessing Module: This module is
responsible for cleaning and preparing the data for
analysis. This could involve tasks such as data
cleaning, feature extraction, and data
transformation.
11. Natural Language Processing (NLP) Module:
This module is responsible for analyzing text data,
such as social media posts and news articles, to
extract relevant information about missing
persons. This could include techniques such as
named entity recognition and sentiment analysis.
12. Image and Video Analysis Module: This
module is responsible for analyzing images and
video footage to identify missing persons or
potential leads. This could include techniques such
as object detection, face recognition, and facial
expression analysis.
13. Machine Learning Module: This module is
responsible for building and training machine
learning models to identify patterns and potential
leads in the data. This could include techniques
such as clustering, classification, and regression.
14. Geospatial Analysis Module: This module is
responsible for analyzing location data, such as
GPS data or crime incident reports, to identify
potential areas where missing persons may be
located. This could include techniques such as heat
mapping and spatial clustering.
15. Decision Support System Module: This module
is responsible for providing decision support to law
enforcement agencies and other stakeholders
based on the results of the analysis. This could
include generating reports, visualizations, and
recommendations based on the data.
16. User Interface Module: This module is
responsible for providing a user-friendly interface
for law enforcement agencies and other
stakeholders to access the results of the analysis
and make informed decisions.
5.3. Algorithms:
The k-Nearest Neighbors (kNN) algorithm is a machine
learning algorithm that could potentially be used in a
project that uses AI to find missing persons.
The kNN algorithm is a type of supervised learning
algorithm that can be used for classification or
regression. In the case of finding missing persons, it
could be used for classification, where the algorithm is
trained on a dataset of known missing persons and
non-missing persons to make predictions about new
cases.
Here's how the kNN algorithm works:
The algorithm is trained on a labeled dataset of missing

persons and non-missing persons. Each data point in
the dataset includes features such as age, gender, last
known location, and other relevant information.
When a new case of a missing person is reported, the

algorithm analyzes the features of the case and
calculates the distance between the new case and the
existing cases in the training dataset.
The kNN algorithm then selects the k nearest cases

based on their distance to the new case, where k is a
user-defined parameter.
The algorithm classifies the new case based on the
majority label of the k nearest cases. For example, if
the majority of the k nearest cases are labeled as
missing persons, the algorithm would classify the new
case as a missing person.
The kNN algorithm is relatively simple and easy to

implement, and it can be effective for classification
tasks with small to medium-sized datasets. However, it
may not perform as well with larger datasets or when
the features are not well-defined or informative.
Note that the kNN algorithm is just one of many

machine learning algorithms that could potentially be
used in a project that uses AI to find missing persons.
The choice of algorithm would depend on the specific
requirements of the project and the type of data that is
available.
6.Test Cases
Test Cases for finding missing person using AI:
4. Test case: Using facial recognition technology
Input: A set of images of the missing person and a

database of images
Output: A match of the missing person's image with
one of the images in the database
5. Test case: Using natural language processing
Input: A description of the missing person's physical

appearance, last known location, and any other
relevant details
description
6. Test case: Using location tracking

database of locations
location data
4. Test case: Using social media monitoring

social media activity
Output: A list of potential matches based on the social
media data
5.Test case: Using machine learning algorithms

information
Output: A list of potential matches based on patterns
and correlations in the data
1.Using facial recognition technology: This test case
involves using a facial recognition algorithm to
compare the missing person's image with a database of
images. The algorithm would analyze the features of
the person's face and compare them to other images in
the database to identify potential matches. This test
case requires a high-quality image of the missing
person and a database of images that includes
potential matches.
2. Using natural language processing: This test case

involves using natural language processing algorithms
to analyze a description of the missing person's
physical appearance, last known location, and any
other relevant details. The algorithm would analyze the
description and compare it to a database of
descriptions to identify potential matches. This test
case requires a detailed and accurate description of the
missing person and a database of descriptions that
3.Using location tracking: This test case involves using

location data from the missing person's last known
location and any subsequent locations to identify
location data and compare it to a database of locations
to identify potential matches. This test case requires
accurate location data and a database of locations that
4.Using social media monitoring: This test case involves

using social media data related to the missing person,
including posts, messages, and other data, to identify
social media data and compare it to a database of
social media activity to identify potential matches. This
test case requires access to the missing person's social
media accounts or public data related to the missing
person, as well as a database of social media activity
that includes potential matches.
5.Using machine learning algorithms: This test case

involves using machine learning algorithms to analyze a
variety of data sources, including images, location data,
social media data, and any other relevant information,
to identify potential matches. The algorithm would
analyze patterns and correlations in the data to identify
potential matches. This test case requires a diverse set
of data sources and a database of potential matches
that includes a variety of data types. It also requires
training the machine learning algorithm with labeled
data to improve accuracy.
6.Using facial recognition technology:
Input: A high-quality image of the missing person and a

database of images.
Algorithm: The algorithm would use a neural network
to analyze the features of the missing person's face and
compare them to other images in the database to
identify potential matches. It would use deep learning
techniques to learn the patterns and features that
distinguish different faces.
Output: The output would be a match of the missing
person's image with one of the images in the database.
7.Using natural language processing:
Input: A detailed and accurate description of the

missing person's physical appearance, last known
location, and any other relevant details, as well as a
database of descriptions.
processing techniques to analyze the description and
identify key features, such as height, weight, hair color,
and clothing. It would compare these features to other
descriptions in the database to identify potential
matches.
matches based on the description.
8.Using location tracking:

database of locations.
Algorithm: The algorithm would use machine learning
techniques to analyze the location data and identify
patterns and correlations. It would compare the
location data to other locations in the database to
identify potential matches.
matches based on the location data.
9.Using social media monitoring:

social media activity.
processing techniques to analyze the social media data
and identify key features, such as interests, activities,
and connections. It would compare these features to
other social media data in the database to identify
potential matches.
matches based on the social media data.
10.Using machine learning algorithms:

information, as well as a database of potential
matches.
Algorithm: The algorithm would use a combination of
machine learning techniques, such as neural networks,
decision trees, and clustering, to analyze the data and
identify patterns and correlations. It would use
supervised learning to train the algorithm with labeled
data and unsupervised learning to identify patterns in
unlabeled data.
matches based on the data sources and the patterns
and correlations identified by the algorithm.
7. Results
8. Sample Code
9. Conclusion
In conclusion, finding missing persons using AI involves
using a combination of techniques such as facial
recognition, natural language processing, location
tracking, social media monitoring, and machine
learning. These techniques can be used to analyze
various data sources and identify potential matches
based on patterns and correlations in the data. The
accuracy of the results depends on the quality and
diversity of the data sources, the algorithms used, and
the availability of relevant databases. While AI can
assist in finding missing persons, it's important to note
that it cannot replace human efforts such as physical
searches and investigations. It can, however, serve as a
valuable tool for law enforcement agencies and search
and rescue teams in their efforts to locate missing
persons.
Here are some additional points to consider regarding

finding missing persons using AI:
Ethical considerations: AI systems used for finding

missing persons must be designed with ethical
considerations in mind. This includes protecting the
privacy of individuals, avoiding bias and discrimination,
and ensuring transparency and accountability.
Integration with existing systems: AI systems for finding

missing persons should be integrated with existing
databases and systems used by law enforcement
agencies and search and rescue teams. This can help
improve the accuracy and efficiency of the search
process.
Real-time monitoring: AI systems can be used for real-

time monitoring of social media and other data sources
to identify potential leads and updates related to the
missing person. This can help search and rescue teams
stay up-to-date with the latest information.
Crowdsourcing: AI systems can also be used to analyze

crowdsourced data such as photos and videos from the
public to identify potential leads and sightings of the
missing person. This can help expand the search
beyond traditional sources of data.
Continuous improvement: AI systems for finding

missing persons should be continuously improved and
updated to ensure they remain accurate and effective.
This includes incorporating new data sources, updating
algorithms, and refining the training data used to
improve the accuracy of the system.
Overall, finding missing persons using AI is a complex

and multi-faceted process that requires careful
consideration of various factors. AI can serve as a
powerful tool in the search process, but it should be
used in conjunction with other human efforts and with
a strong emphasis on ethics, transparency, and
accountability.

Finding Missing Persons with AI

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Finding Missing Persons with AI

Uploaded by

Copyright:

Available Formats

1.

Face recognition is a biometric-based technology

In the world, a countless number of people are missing

1.2 . Project Overview:

The aim of this project is to develop a system that uses

2. Technologies and Techniques

2.1 . Natural Language Processing (NLP):

2.2 Image Recognition:

Machine Learning: Machine learning algorithms

2.3 .Data Mining:

Graph Databases: Graph databases can be used

User Interface Design: A user interface can be

Cloud Computing: Cloud computing can be used

3.1 Problem Statement:

3.2 Scope of the System:

The proposed system contains the following Modules:

3.3 Functional Requirements:

A user interface to present the results to users

4.1. System Design:

4.2. UseCase Diagram:

Note that this is just an example use case diagram, and

4.3. Class Diagram:

Finally, the Algorithm class represents the various AI

4.5. Activity Diagram:

4.6. Deployment Diagram:

The Results System is responsible for displaying the

When the desktop application registers about

Here's how the kNN algorithm works:

The algorithm is trained on a labeled dataset of missing

When a new case of a missing person is reported, the

The kNN algorithm then selects the k nearest cases

The kNN algorithm is relatively simple and easy to

Note that the kNN algorithm is just one of many

Input: A set of images of the missing person and a

2. Test case: Using natural language processing

Input: A description of the missing person's physical

3. Test case: Using location tracking

Input: GPS data from the missing person's last known

4. Test case: Using social media monitoring

Input: Social media posts, messages, and other data

5.Test case: Using machine learning algorithms

Input: A variety of data sources, including images,

2. Using natural language processing: This test case

3.Using location tracking: This test case involves using

4.Using social media monitoring: This test case involves

5.Using machine learning algorithms: This test case

6.Using facial recognition technology:

Input: A high-quality image of the missing person and a

7.Using natural language processing:

Input: A detailed and accurate description of the

8.Using location tracking:

Input: GPS data from the missing person's last known

9.Using social media monitoring:

Input: Social media posts, messages, and other data

10.Using machine learning algorithms:

Face recognition is a biometric-based technology

In the world, a countless number of people are missing

1.2 . Project Overview:

The aim of this project is to develop a system that uses

6. Technologies and Techniques

6.1 . Natural Language Processing (NLP):

6.2 Image Recognition:

Machine Learning: Machine learning algorithms