You are on page 1of 10

Team Name : Tech Warriors

Problem Statement : AI - Face Detection


BRIEF ABOUT THE IDEA:
Face detection , a computer technology is a solution we bring up to identify
MISSING PERSON. It is used in a lot of biometrics where we detect unique
things on the face like moles , scars , dark spots.

1. Though we have inbuilt models like Tensor Flow and Open CV, we are
implementing DNN so that unique features will be extracted more accurately.
2. Here we use EAR as a unique feature to recognize a person even if they
wear a mask. In addition to particular angles, we use a single ear to
recognize a person.
3. We cannot expect higher quality videos every time . So we trained our
model to detect objects in low quality videos too.
OPPORTUNITIES:
How different is it from any other existing ideas out there?
EAR BASED RECOGNITION:
We detect person based on
Ear besides him/her wearing 7 ANGLE TECHNIQUE:
a mask and tilted angles. Creating higher chances to
detect the person from
TAMPERING DETECTION: different angles.
Detecting an active attempt
to compromise the device
integrity or associated data. PUBLIC DISRUPTION:
Incase of any public
disturbances our DNN model
CLUSTERING OF identifies them and reports to
CAMERAS: the nearby police station.
Way of grouping data points
from different devices
(cameras).
LIST OF FEATURES OFFERED BY THE SOLUTION:

OPTICAL CLUSTERING OF
CHARACTER CAMERAS
RECOGNITION
FEATURES
OFFERED

IDENTIFICATION
OF MISSING
PERSON BESIDE
DETECTION OF WEARING A MASK
TAMPERS

BLOCKCHAIN FOR
FIR FILING
WORKING FLOW:
1.DATA COLLECTION 3.LOGISTIC REGRESSION
Collection of face images of It is used for classification of
the missing person. images.

2.DNN MODEL 4.PIXEL SUBTRACTION


Low weighted user The digital numerical value of one
defined model with high pixel or whole image is subtracted
accuracy. from previous image .It is used for
Tamper detection.
START

Training our model

Collecting the image


of the missing person
True If False
accuracy
>85%
Convert the image into 7
different angles Person Person not
recognised recognised

Augument the image using


Missing
tensor flow
person found

Import libraries
STOP
BUSINESS LOGIC OF THE SOLUTION:
1.SURVEILLANCE: 2.SECURITY PURPOSES:
Object Tracking – To monitor in and out of the traversing object. We use this model in houses , public
Frequently used in jewellery shops for monitoring customers. places like railway stations to avoid the
Keywords: 0 – Monitoring the entry of regular customers. occurrence of crime.
1 – Monitoring the exit of regular customers.
2 – Monitoring the normal customers.

3.VIDEO ANALYSIS:
 Used to find missing person among the crowd.
 Allows athletes to take a look at their performance to improve their skill.
 Monitors the whole action in cricket in real time/slow motion , forward/backward.
TECHNOLOGIES USED:
ML , Deep Learning,
OpenCV, Keras,
• We are using DNN model to recognize the person’s face and
logistic regression used for classification and Open CV, Keras
Tensor Flow are used for the computer vision to detect the person.

Ethereum Block Chain,


MetaMask
• It is a decentralized network where we can upload the
details of FIR so that no one can change the details.

CCTV or Video Capture


Device
• It’s a hardware component where we need these things
to detect people.

High Capacity computers • We need to have GPU inorder to work in a Deep


learning model.
ESTIMATED COST OF/AFTER IMPLEMENTING THE
SOLUTION
COMPONENTS PRICE

CCTV cameras 8000/- per CCTV

High tech computers 1,00,000/- per computer

Business intelligence , ML 50,000/-


engineer
Block chain gas fee (Average of 9$
the blocks created )

 The estimated price of this security system is Rs. 2,00,000.


THANK YOU!

You might also like