undefined. Team Members Humaira Zaman (SP21-BCS-024) Muhammad Mubeen Bhatti (SP21-BCS-025). 

Scene 1

Team Members
 Humaira Zaman (SP21-BCS-024) Muhammad Mubeen Bhatti (SP21-BCS-025)

Department of Computer Science COMSATS University Islamabad, Wah Campus

Presentation Outline. Introduction Existing System & Our Contributions Brief Description Technologies & Tools Project Schedule Questions and Answers. 

Scene 2

Introduction Existing System & Our Contributions Brief Description Technologies & Tools Project Schedule Questions and Answers

Department of Computer Science, CUI Wah Campus

undefined. Introduction. Project Track: Mobile Application with Generative AI. 

Scene 3

Project Track: Mobile Application with Generative AI

Brief Introduction of the proposed idea: An app-based platform to convert sign language into text using generative AI concepts. We will make our own dataset to train the model. Trained model will detect and recognize the sign and transformers will be used to convert those signs into text.

undefined. Introduction [Contd.]. Objectives of the proposed idea: Trained model with high accuracy. Mobile App for seamless interaction. Real-time translation of sign language gestures into text or spoken. Applications of the proposed idea: This application will be used by the deaf and normal to make connection between each other.. 

Scene 4

Objectives of the proposed idea: Trained model with high accuracy. Mobile App for seamless interaction. Real-time translation of sign language gestures into text or spoken.
 Applications of the proposed idea: This application will be used by  the deaf and normal to make connection between each other.

Existing Systems & Our Contribution. (5). Department of Computer Science, CUI Wah Campus. 

Scene 5

Sr. No Title Dataset Algorithms/Libraries Features Accuracy 1 Real Time Sign Language Detection 5 gestures(ASL): “Hello”, “Yes”, “No”, “Thank you” and “I Love You” OpenCV, TensorFlow, Convolutional Neural Network (ConvNet/CNN) Realtime Sign Detection 80-90% 2 Real Time Identification of ASL for Deaf and Dumb Community American Sign language (ASL) (A-Z)(0-9) OpenCV, CNN Voice Conversion(Text-voice) 80-above% 3 Special Child Trainer, Deaf and Dumb Application American Sign Language Convolutional Neural Network (CNN) Learning and Testing 80-90%

Existing Systems & Our Contribution. (6). Department of Computer Science, CUI Wah Campus. 

Scene 6

Sr. No Title Dataset Algorithms/Libraries Features Accuracy 4 Detection and Recognition of Sign Language and it’s Translation to Urdu Urdu Sign Language (Alphabets) TensorFlow, OpenCV Real-Time Recognition through Webcam 20-above% 5 Deaf sign Translation By GenAI Pakistan Sign Language(PSL) OpenCV, Transformer Sign to text/audio, text to sign  Our proposed idea

[Audio] Module(s) of the system are:
Data Collection
Data Preprocessing Model Training Real Time Translation App Development Functionalities of each modules are:
Data Collection:
It contains Pakistani Deaf signs and will be collected through OpenCV.
Data will be in form of images and videos.. 

Scene 7

Module(s) of the system: Data Collection Data Preprocessing  Model Training  Real Time Translation  App Development
 Functionalities of each module:

Data Collection: It contains Pakistani Deaf signs and will be collected through OpenCV. Data will be in form of images and videos.

[Audio] Data Preprocessing:
Extract features of images/videos and make noiseless for the training.
Model Training:
Transformer Library will be used.
Real Time Translation:
Google A-P-I will be used for conversion.
App Development:
App will be built using Dart(with Flutter).. 

Scene 8

Data Preprocessing: Extract features of images/videos and make noiseless for the training. Model Training: Transformer Library will be used. Real Time Translation: Google API will be used for conversion. App Development: App will be built using Dart(with Flutter).

Brief Description [Contd.]. (9). Department of Computer Science, CUI Wah Campus. 

Scene 9

Members Workload justification Humaira Zaman (SP21-BCS-024)
 Data Preprocessing Model Training Testing & Validation Muhammad Mubeen Bhatti (SP21-BCS-025) Data Collection Real time Translation Mobile App Development

Member-wise workload justification with respect to modules/functionalities:

Jupyter Notebook. Google Colab. Python. Technologies & Tools. 

Scene 10

Project Schedule. (11). Department of Computer Science, CUI Wah Campus. 

Scene 11

Sr. No. Semester  Milestone Deliverables Deadline 1 VII Proposal  Document Mar, 2024 2 VII Software Requirement Specification & Design Specification Document Apr, 2024 3 VII Implementation of 25% of the Functional Requirements Working Software Jun, 2024 4 VIII Implementation of 90-100% of the Functional Requirements + FYP Report Working Software + Document 8th Semester 5 VIII Final Project Presentation + Final FYP Report Complete Software + Document 8th Semester

Questions/Answers. (12). Department of Computer Science, CUI Wah Campus. 

Scene 12

A person and person standing next to each other

Description automatically generated

iHearU: Deaf Sign Translation by GenAI

Scene 0

Module(s) of the system are:
Data Collection
Data Preprocessing 
Model Training 
Real Time Translation 
App Development

Functionalities of each modules are:
Data Collection:
It contains Pakistani Deaf signs and will be collected through OpenCV.
Data will be in form of images and videos.

avatar

Data Preprocessing:
Extract features of images/videos and make noiseless for the training.
Model Training:
Transformer Library will be used.
Real Time Translation:
Google A-P-I will be used for conversion.
App Development:
App will be built using Dart(with Flutter).

iHearU: Deaf Sign Translation by GenAI

Scene 1 (0s)

Scene 2 (9s)

Scene 3 (14s)

Scene 4 (39s)

Scene 5 (59s)

Scene 6 (1m 4s)

Scene 7 (1m 9s)

Scene 8 (1m 13s)

Scene 9 (1m 37s)

Scene 10 (1m 41s)

Scene 11 (1m 46s)

Scene 12 (1m 51s)