An Intelligence Security System
for Women
-K M Aishwarya
-S R Prathiksha
-Shubhangi Agarwal
Why this?
In the month of September, as a part of Hike India’s and
Dexter Talent Lab’s collaborative & innovative approach
to increase women safety in our society, we participated
in a Hackathon organized for women that was aimed at
building a software product that helps increase safety for
women in India.
Being fortunate enough to secure the 3rd position in the
hackathon, we got a chance to visit the Hike Messenger’s
HQ in New Delhi, India; and present our protocol to the
CEO & Founder of Hike: Kavin Bharti Mittal.
Overview
Enclosed within these slides is a brief description of:
• The problem statement we tried to solve
• Our approach towards it
• The solution we came up with.
Problem Statement
Build a security system for women that is
completely automated and requires no human
interference whatsoever.
Approach
The system should be fool proof and
independent of any mobile or web application.
So that safety is ensured even if the victim is not
in possession of a phone at the time of the
crime.
Solution
• We built a MATLAB-web interface wherein:
– An arrangement of a bi-camera setup (2 cameras)
– One placed at a higher altitude than the other.
– Camera 1: Captures image. Detects the number of people in the image.
– If more than 1 person detected, signal sent to Camera 2.
– Camera 2: Captures image. Gender recognition performed based on facial
features.
– If even one female is recognized, Facial Expression Recognition is
performed.
– On detecting emotions of fear, disgust or sadness, Camera 2 sends off an
alarm to the web server.
– Backend code retrieves latitude and longitude of the camera where threat
is detected.
– Alert SMS’s and emails sent to the nearby police stations by web server.
I. MATLAB Simulation
• Version used: R2009a
• Implemented for: Facial Expression
Recognition (Eigen Face Recognition Method)
Why use MATLAB?
– In built toolboxes: Image Processing toolbox,
Artificial Neural Networks toolbox, audio-video
processing toolbox
– Pre-processing is way easier!
– A vast base for simulation of both hardware and
software components.
– Incorporates Linear Algebra for all mathematical
calculations.
Mood Detection
• The 7 basic moods detected on human face in
MATLAB:
– Happiness
– Sadness
– Fear
– Disgust
– Angry
– Surprise
– Neutral
Steps
• Capture an image
• Pre-processing of input image
• Perform PCA (Principal Component Analysis),
identify and concentrate on the skin-area of the
face.
• Perform Facial Expression Recognition using Eigen
Face recognition.
• Display output image with the expression label of
the matched image from the train database.
Step 1 - Preprocessing
– Denoising the image (removes unwanted
distortions that are seen as noise)
– Feature extraction (focuses on facial features)
– Cropping (background is cropped)
– Recognizing and focusing on the face area
Step 2 - PCA
Few instances of usage of PCA:
– Dimensionality reduction (used in this case)
– Pattern Recognition
– Data Reduction (audio/video/image etc.)
PCA is a vast topic in itself, which cannot be covered within these slides
alone. Hence only its areas of application have been mentioned here.
Step 3 - Facial Expression Recognition
• Given a prior database of images with each one
of them labeled with a facial expression, an input
captured image has to be pre-processed,
compressed and matched with one of the images
in the training set.
• Once recognized, the label assigned to the image
identified from the training set is assigned to
input image.
• Hence the facial expression of the captured
image is displayed, based on which an alarm is
set off.
II. Developing the User
Interface
UI – AngularJS
• Framework for dynamic web apps.
• Mainly used for single page applications.
• A product of Google©.
• Features –
– MVVM(Model-View-ViewModel)
– 2 Way Data Binding
– Filters and dynamic templating.
Working
• Backend code sends the latitude and longitude
of the place where a threat gets detected.
• Reverse geocoding the latitude and longitude
attributes and displaying of the location on
Google maps, by creating markers.
• Thus, location is sent to helping centers and
police stations nearby.
Why AngularJS?
• Dynamically fetches results and updates view.
• Pairs with AJAX for amazing speeds.
• Can be easily incorporated with iOs and
Android.
Emergency Alert Service (EAS)
 Instant email alerts sent to all the Durga Centers (women help
centers), police stations and the volunteers nearby.
 Provision for voluntary signup as volunteers on the website made
available.
 Send Text Messages, Hike messages to volunteers according to
their preferences.
 Finding out the availability of volunteers as to whether they are in
town or not.
 In cases of the volunteer being out of reach, the message alerts
can be disabled.
 Provision of extra services (like automated calls) parallely with the
message notifications.
III. Tech Stack
Spring framework.
J2EE
REST APIs
MySQL
• Why APIs?
For reusability purposes.
• Why J2EE and Spring?
Tends to impose as little constraints as possible.
• Why MySQL?
Faster and more reliable.
Thank you!
We hope you go ahead and contribute and make
this world a better place to be in! 
Any questions?
Drop a mail to: aish3095@gmail.com

An Intelligence Security System for Women

  • 1.
    An Intelligence SecuritySystem for Women -K M Aishwarya -S R Prathiksha -Shubhangi Agarwal
  • 2.
    Why this? In themonth of September, as a part of Hike India’s and Dexter Talent Lab’s collaborative & innovative approach to increase women safety in our society, we participated in a Hackathon organized for women that was aimed at building a software product that helps increase safety for women in India. Being fortunate enough to secure the 3rd position in the hackathon, we got a chance to visit the Hike Messenger’s HQ in New Delhi, India; and present our protocol to the CEO & Founder of Hike: Kavin Bharti Mittal.
  • 3.
    Overview Enclosed within theseslides is a brief description of: • The problem statement we tried to solve • Our approach towards it • The solution we came up with.
  • 4.
    Problem Statement Build asecurity system for women that is completely automated and requires no human interference whatsoever.
  • 5.
    Approach The system shouldbe fool proof and independent of any mobile or web application. So that safety is ensured even if the victim is not in possession of a phone at the time of the crime.
  • 6.
    Solution • We builta MATLAB-web interface wherein: – An arrangement of a bi-camera setup (2 cameras) – One placed at a higher altitude than the other. – Camera 1: Captures image. Detects the number of people in the image. – If more than 1 person detected, signal sent to Camera 2. – Camera 2: Captures image. Gender recognition performed based on facial features. – If even one female is recognized, Facial Expression Recognition is performed. – On detecting emotions of fear, disgust or sadness, Camera 2 sends off an alarm to the web server. – Backend code retrieves latitude and longitude of the camera where threat is detected. – Alert SMS’s and emails sent to the nearby police stations by web server.
  • 7.
    I. MATLAB Simulation •Version used: R2009a • Implemented for: Facial Expression Recognition (Eigen Face Recognition Method)
  • 8.
    Why use MATLAB? –In built toolboxes: Image Processing toolbox, Artificial Neural Networks toolbox, audio-video processing toolbox – Pre-processing is way easier! – A vast base for simulation of both hardware and software components. – Incorporates Linear Algebra for all mathematical calculations.
  • 9.
    Mood Detection • The7 basic moods detected on human face in MATLAB: – Happiness – Sadness – Fear – Disgust – Angry – Surprise – Neutral
  • 10.
    Steps • Capture animage • Pre-processing of input image • Perform PCA (Principal Component Analysis), identify and concentrate on the skin-area of the face. • Perform Facial Expression Recognition using Eigen Face recognition. • Display output image with the expression label of the matched image from the train database.
  • 11.
    Step 1 -Preprocessing – Denoising the image (removes unwanted distortions that are seen as noise) – Feature extraction (focuses on facial features) – Cropping (background is cropped) – Recognizing and focusing on the face area
  • 12.
    Step 2 -PCA Few instances of usage of PCA: – Dimensionality reduction (used in this case) – Pattern Recognition – Data Reduction (audio/video/image etc.) PCA is a vast topic in itself, which cannot be covered within these slides alone. Hence only its areas of application have been mentioned here.
  • 13.
    Step 3 -Facial Expression Recognition • Given a prior database of images with each one of them labeled with a facial expression, an input captured image has to be pre-processed, compressed and matched with one of the images in the training set. • Once recognized, the label assigned to the image identified from the training set is assigned to input image. • Hence the facial expression of the captured image is displayed, based on which an alarm is set off.
  • 14.
    II. Developing theUser Interface
  • 15.
    UI – AngularJS •Framework for dynamic web apps. • Mainly used for single page applications. • A product of Google©. • Features – – MVVM(Model-View-ViewModel) – 2 Way Data Binding – Filters and dynamic templating.
  • 16.
    Working • Backend codesends the latitude and longitude of the place where a threat gets detected. • Reverse geocoding the latitude and longitude attributes and displaying of the location on Google maps, by creating markers. • Thus, location is sent to helping centers and police stations nearby.
  • 17.
    Why AngularJS? • Dynamicallyfetches results and updates view. • Pairs with AJAX for amazing speeds. • Can be easily incorporated with iOs and Android.
  • 18.
    Emergency Alert Service(EAS)  Instant email alerts sent to all the Durga Centers (women help centers), police stations and the volunteers nearby.  Provision for voluntary signup as volunteers on the website made available.  Send Text Messages, Hike messages to volunteers according to their preferences.  Finding out the availability of volunteers as to whether they are in town or not.  In cases of the volunteer being out of reach, the message alerts can be disabled.  Provision of extra services (like automated calls) parallely with the message notifications.
  • 19.
    III. Tech Stack Springframework. J2EE REST APIs MySQL • Why APIs? For reusability purposes. • Why J2EE and Spring? Tends to impose as little constraints as possible. • Why MySQL? Faster and more reliable.
  • 20.
    Thank you! We hopeyou go ahead and contribute and make this world a better place to be in! 
  • 21.
    Any questions? Drop amail to: aish3095@gmail.com