You are on page 1of 18

CRIMINAL IDENTITY DETECTOR

Project Report
Project Work Phase 2 (ECS 899)
Degree
BACHELOR OF TECHNOLOGY (CSE)

PROJECT GUIDE: SUBMITTED BY:


Prof. Ankur Sisodia Samyak Jain(TCA1709030)
Samyak Jain(TCA1709021)

June, 2021

FACULTY OF ENGINEERING & COMPUTING SCIENCES


TEERTHANKER MAHAVEER UNIVERSITY, MORADABAD
DECLARATION

We here by declare that this Project Report titled Criminal Identity Detector
submitted by us and approved by our project guide, Faculty of Engineering &
Computing Sciences. Teerthanker Mahaveer University, Moradabad, is a bonafide
work undertaken by us and it is not submitted to any other University or
Institution for the award of any degree diploma / certificate or published any time
before.

Project ID : ECSP-13

Student Name: Samyak Jain Signature

Student Name: Samyak Jain Signature

Project Guide : Ankur Sisodia Signature


TMU-FOE&CS Version 5.0 T003A-Project Report

Table of Contents
1 PROJECT TITLE............................................................................................................................................ 4
2 PROBLEM STATEMENT................................................................................................................................ 4
3 PROJECT DESCRIPTION................................................................................................................................ 4
3.1 SCOPE OF THE WORK..........................................................................................................................................4
3.2 PROJECT MODULES..........................................................................................................................................4
3.3 CONTEXT DIAGRAM (HIGH LEVEL)......................................................................................................................4
4 IMPLEMENTATION METHODOLOGY............................................................................................................ 4
5 TECHNOLOGIES TO BE USED........................................................................................................................ 4
5.1 SOFTWARE PLATFORM.....................................................................................................................................4
5.2 HARDWARE PLATFORM....................................................................................................................................5
5.3 TOOLS, IF ANY................................................................................................................................................ 5
6 ADVANTAGES OF THIS PROJECT.................................................................................................................. 5
7 ASSUMPTIONS, IF ANY................................................................................................................................ 5
8 FUTURE SCOPE AND FURTHER ENHANCEMENT OF THE PROJECT.................................................................5
9 PROJECT REPOSITORY LOCATION................................................................................................................ 5
10 DEFINITIONS, ACRONYMS, AND ABBREVIATIONS........................................................................................ 6
11 CONCLUSION.............................................................................................................................................. 6
12 REFERENCES............................................................................................................................................... 7

Appendix
A: Data Flow Diagram (DFD)
B: Entity Relationship Diagram (ERD)
C: Use Case Diagram (UCD)
D: Data Dictionary (DD)
E: Screen Shots

Project Title: CID Page 3 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

1 Project Title
Criminal Identity Detector

2 Problem Statement
Hand-drawn forensic sketch based on witness descriptions and searching of the criminal based
on sketch are both time consuming method. We can automate this work by using deep learning
algorithms which can identify criminal quickly and accurately.

3 Project Description
CID is a application that helps to identify criminal based on eye witness information. It draws
the image of criminal based on user voice input and search the images from the database. It
replaces the work of forensic artist and saves the time of searching the criminal. It does this
quickly and accurately using deep learning algorithms.

3.1 Scope of the Work


It will be used by investigation teams because searching the criminal by uncleared image is very
time consuming task. It need images of all the people of India which can be possible if this
project is hand over to government. Peoples facial features change at different phases of life so
it need image database creation for one time and will update these image using AI aging
algorithm.

3.2 Project Modules

1) Search Image Module


• 1.1 Description
This module allows user to saves their time in searching the criminal.

 1.2 Inputs
It requires voice input from user.

• 1.3 Processing
After generating the image it search the image in the database.

• 1.4 Outputs
Searched Images.

2) Delete Image Module


4.1 Description
This module delete a criminal or people images who died .

4.2 Inputs
Customer voice input is entered.

Project Title: CID Page 4 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

4.3 Processing

Further Processing is continued if the input is in required format.

4.4 Outputs
Database update successfully.

3) Update Image Module


6.1 Description
This module allows to update the images of people in every two years using AI aging algorithm.

 3.1.2 Inputs
This module requires images of database to apply algorithm.

• 3.1.3 Processing
Provided input is processed and data is updated in the database.

• 3.1.4 Outputs
All the images of databases updated.

3.3 Context Diagram (High Level)

Voice Input
Criminal
Identity
User
Detector

Searched Image Output

4 Implementation Methodology
Step 1- Model and Dataset Building
Models- Stack GAN Stage 1 and Stage 2, Siamese Neural Network.
Dataset- 5000 Image caption with Image pair require for model training.
Step 2- Desktop GUI Building
Tkinter based GUI require to process input and give desire output.

Project Title: CID Page 5 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Step 3- Integration of Trained Models with GUI


Step 4- Database Creation
Storing every person details in JSON file with images in a folder that will be fetch by GUI to
search image that matches with fake image generated by Stack GAN model using Siamese
Neural Network model.

5 Technologies to be used
5.1 Software Platform
a) Front-end
Python Tkinter
b) Back-end
Stack GAN Algorithm
Siamese Neural Network
PyTorch
Speech recognition API

5.2 Hardware Platform


RAM-12GB, CPU, GPU, TPU

5.3 Tools, if any


Google Colab
Google Drive

6 Advantages of this Project


It automates the work of forensic artist and saves the time of searching the criminal.

7 Assumptions, if any
Assuming faces will change with time need to update the images face aging algorithm.

8 Future Scope and further enhancement of the Project


It will be used by forensic teams. Currently, Stack GAN not generating image according to text
but that problem will be resolve by using Attention GAN model and also image can be directly
search by Siamese Neural Network without generating image. If application performing
according to requirement then it need time to time image updating by Face Aging deep learning
algorithm.

Project Title: CID Page 6 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

9 Project Repository Location


S Project Location Verified Verified
# Artifacts (Mention Lab-ID, Server ID, Folder Name etc.) by by Lab
(softcopy) Project In-
Guide Charge
1. Project Lab-3206 Name Name
Synopsis and and
Report (Final Signatur Signatur
Version) e e
2. Project Lab-3206 Name Name
Progress and and
updates Signatur Signatur
e e
3. Project Lab-3206 Name Name
Requirement and and
specification Signatur Signatur
s e e
4. Project Lab-3206 Name Name
Report (Final and and
Version) Signatur Signatur
e e
5. Test Google Colab- “https://colab.research.google.com/drive Name Name
Repository and and
/19Xc_yyFm9JWg86lTuNx2c4QHNXx47rIR?usp=sharing”
Signatur Signatur
e e
6. Project Github- Name Name
Source Code https://github.com/JainSamyak8840/criminal_identity_detect and and
(final or Signatur Signatur
version) with e e
executable

7. Power Point Folder-project synopsis Name Name


Presentation and and
Signatur Signatur
e e

10 Definitions, Acronyms, and Abbreviations

Abbreviation Description

Project Title: CID Page 7 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

CID Criminal Identity Detector


GAN Generative Adversarial Network

11 Conclusion
If we wants to implement this application on large scale then we has to hand over this project to
government then government can collect images at time of voter ID creation because it requires 18+
aged people who are legally consider for crime and after that age very our face gets slow changes. It
requires image updation within two to three year using face-aging GAN algorithm.
It is not for personal use , it will be used professionally by investigation teams.

12 References

S# Reference Details Owner Version Date


1. Project Synopsis ECSP-13 1.0 08-08-20

2. Project Requirements ECSP-13 09-09-20


3. arxiv.org

4. medium.com
5. towarddatascience.com

Project Title: CID Page 8 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Annexure A
Data Flow Diagram (DFD)
(Mandatory)
Level-0

Criminal Identity
Detector

User Voice Input Output


Image

Project Title: CID Outputting Page 9 of 18

Searched image
TMU-FOE&CS Version 5.0 T003A-Project Report

Data Flow Diagram (DFD)


(Mandatory)
Level-1

Speech Siamese
Recognition STACK GAN Neural Network
API Model Model

Stage 1 Stage 2 Stage 3


Voice to text Text to Image
Image
Search

User Voice Input


Output
Image DB Image

Project Title: CID Page 10 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Annexure B
Entity-Relationship Diagram (ERD)
(Mandatory)

Voice to Text
Inputting
User
Voice
Converter

Passing
Text

Databases Text to Image


Translator

Fetching
Images

Output Image Passing


Image

Outputting Image to Image

Searched image Comparison

Project Title: CID Page 11 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Annexure D
Data Dictionary (DD)
(Mandatory)

Example:
User Table (USR)
Fields Data type Description
USR-Voice Input Text Criminal face description

Supplier Table (SUPP)


Fields Data type Description
SUPP-ID Number Supplier ID
SUPP-Name Text Supplier Name
SUPP-Address Text Supplier Address
SUPP-Contact Number Supplier Contact
SUPP-Crime-Count Number Crime Count

Project Title: CID Outputting Page 12 of 18


Outputting
Searched image
Searched image
TMU-FOE&CS Version 5.0 T003A-Project Report

Annexure E
Screen Shots
Testing-1
Home Page:
Here user gives voice input by pressing microphone button that converts to text by speech
recognition API.

Project Title: CID Page 13 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Stage 2 Page:
This stage comes after pressing draw button on home page. It contains the image generated
from text by Stack GAN model.

Project Title: CID Page 14 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Stage 3 Page:
This stage comes after pressing search button. Below image is searched by Siamese neural
network model by comparing the above generated image with the images in database.

Project Title: CID Page 15 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Testing-2
Stage 1 Page:

Project Title: CID Page 16 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Stage 2 Page:

Project Title: CID Page 17 of 18


TMU-FOE&CS Version 5.0 T003A-Project Report

Stage 3 Page:

Project Title: CID Page 18 of 18

You might also like