Scribd
Upload
15 views

Wearable Translation Device for Staff

The document presents a wearable device designed to eliminate language barriers in medical settings through real-time translation between medical staff and patients. It outlines the device's components, operational steps, current progress, and a SWOT analysis highlighting its strengths, weaknesses, opportunities, and threats. Additionally, it identifies the target audience, emphasizing the significant Tamil-speaking population in Sri Lanka.

Uploaded by

en91418
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
15 views

Wearable Translation Device for Staff

The document presents a wearable device designed to eliminate language barriers in medical settings through real-time translation between medical staff and patients. It outlines the device's components, operational steps, current progress, and a SWOT analysis highlighting its strengths, weaknesses, opportunities, and threats. Additionally, it identifies the target audience, emphasizing the significant Tamil-speaking population in Sri Lanka.

Uploaded by

en91418
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

FINAL PRESENTATION

anslation Dev
Tr ice
ble fo
ra r
GROUP 07 - SUB GROUP 02

S
a ta
e
W

ff
What is this?
This is a wearable device which facilitates zero
language barriers between medical staff and
patients to improve understanding better
.
This device has two components.

The wearable device which is responsible


for getting extracting the sentences and
sending them to cloud.

The cloud server where the translation


happens and data is stored if necessary.

GROUP 07 - SUB GROUP 02


Step 01

How this work?


Step 01 Step 02 Step 03 Step 04 Step 05

The input Using ML Saved file The TTS


The
voice of the based STT will be model will
translated
patient/doct model, send to the convert the
cloud
files will be text based
ors are words are
server to sent back output into
extracted extracted
translate to ESP32 voice based
using and saved
into the microcontr output and
microphone in a file
further desired oller. discarded
module.
language. from the
power.
speaker.

GROUP 07 - SUB GROUP 02


Current progress
PROJECT PLAN

Understand key EMBEDDED DESIGN


milestone and propose
solution
configure and design SOFTWARE SIMULATIONS
embedded system
Code the machine
learning part and TESTING AND DEPLOYMENT
connecting with google
translator Ensuring the
functionality and
usability of the
device.

GROUP 07 - SUB GROUP 02


Developments So Far
Machnine learning base ML model
Designing of the STT ML model and the TTS models for
effective conversion between text and speech

Prototype of the device


A small prototype which shows the working mechanism
of the device itself

Web Server setup


A server to temporarily hold the sending data
and perform the online translation between
languages

GROUP 07 - SUB GROUP 02


Machine learning models

GROUP 07 - SUB GROUP 02


GROUP 07 - SUB GROUP 02
GROUP 07 - SUB GROUP 02
HOW THIS WORKS
This model basically perform several
tasks within the system. They are as
below,

1. Speech-to-Text (STT) Conversion


2. Text Translation
3. Text-to-Speech (TTS) Conversion
4. End-to-End Pipeline

Possible issues:
Computational resource related issues
with smaller controllers like ESP32. Better
to consider using Jetson Nano, Rasberry Pi
4, etc.

GROUP 07 - SUB GROUP 02


Prototype of the device

GROUP 07 - SUB GROUP 02


GROUP 07 - SUB GROUP 02
SWOT analysis for the project
Srengths
1.⁠⁠Provides a novel solution to a long-standing communication barrier in the medical field.
2.⁠⁠Promotes equity in healthcare by bridging language gaps, especially in Tamil-speaking regions.
3.⁠ ⁠Directly addresses a critical issue in areas like Jaffna, Hatton, Nuwara Eliya, and Batticaloa where Tamil-
speaking populations are prominent.
4. ⁠Incorporates advanced real-time translation technology, enhancing the efficiency and accuracy of
communication.
5. ⁠Can potentially be expanded to other fields, such as education and public services, where language barriers
exist.

Weaknesses
1.⁠ ⁠Real-time language translation requires sophisticated algorithms and may face issues with dialects, medical
terminologies, and accents.
2.⁠ ⁠High development and production costs may make the watch less affordable, especially in underfunded
healthcare settings.
3.⁠⁠Resistance from doctors or patients to adopt new technology due to lack of familiarity or trust.
4.⁠ ⁠Reliance on internet connectivity and hardware durability could be a limitation in rural or resource-limited
settings.

GROUP 07 - SUB GROUP 02


SWOT analysis for the project
Opportunities
1.⁠ ⁠Potential for funding or partnership with organizations focused on healthcare and social harmony.

(Government and NGO)

2.⁠⁠Opportunity to expand beyond Sri Lanka to other multilingual countries with similar issues.

3.⁠⁠The project could lead to innovations in AI-powered translation tailored to specific fields like medicine.

4.⁠⁠Improves doctor-patient relationships, reduces misdiagnoses, and fosters ethnic cohesion.

5.⁠⁠Opportunity to add additional features like transcription or data storage for patient records.

Threats
1.⁠⁠Emergence of similar technologies from global tech companies.

2.⁠⁠Risk of sensitive medical information being leaked or misused.

3.⁠⁠Challenges in obtaining approval for medical technology and compliance with local regulations.

4.⁠⁠Misinterpretation of translated content could lead to misunderstandings or mistrust.

5.⁠⁠Rapid changes in technology could render the product obsolete or less competitive.

GROUP 07 - SUB GROUP 02


Practical challenges with the development process
Hardware Limitations of ESP32
Limited Processing Power
Memory Constraints
Storage Limitations
Alternatives: NVIDIA Jetson Nano, Raspberry Pi 4

Real-Time Processing
Latency

Power Consumption
Battery Life
Heat Dissipation

Audio Input and Output


Microphone Quality
Speaker Quality

GROUP 07 - SUB GROUP 02


Identification of the target audience
District Population(2012) Sinhala Speaking Sinhala % Tamil Speaking Tamil %

Colombo 2,324,349 1,778,971 76.54% 545,378 23.46%

Gampaha 2,304,833 2,086,469 90.51% 218,364 9.49%

Kalutara 1,221,948 1,060,107 86.75% 161,841 13.25%

Kandy 1,375,382 1,023,488 74.44% 351,894 25.56%

Matale 484,531 391,305 80.78% 93,226 19.22%

Nuwara
711,644 282,053 39.64% 429,591 60.36%
Eliya

Galle 1,063,334 1,003,722 94.39% 59,612 5.61%

Matara 814,048 767,580 94.30% 46,468 5.70%

Hambantota 599,903 582,301 97.06% 17,602 2.94%

Jaffna 583,882 2,284 0.39% 581,598 99.61%

Mannar 99,570 2,305 2.32% 97,265 97.68%

Vavuniya 172,115 17,138 9.96% 154,977 90.04%

Mullaitivu 92,238 8,927 9.68% 83,311 90.32%

GROUP 07 - SUB GROUP 02


Identification of the target audience
Kilinochchi 113,510 1,331 1.17% 112,179 98.83%

Batticaloa 526,567 6,797 1.29% 519,770 98.71%

Ampara 649,402 252,458 38.88% 396,944 61.12%

Trincomalee 379,541 101,483 26.74% 278,058 73.26%

Kurunegala 1,618,465 1,479,863 91.43% 138,602 8.57%

Puttalam 762,396 560,838 73.55% 201,558 26.45%

Anuradhapura 860,575 782,808 90.99% 77,767 9.01%

Polonnaruwa 406,088 368,197 90.69% 37,891 9.31%

Badulla 815,405 595,321 73.01% 220,084 26.99%

Moneragala 451,058 428,104 94.91% 22,954 5.09%

Ratnapura 1,088,007 947,811 87.12% 140,196 12.88%

Kegalle 840,648 718,369 85.45% 122,279 14.55%

Sri Lanka 20,359,439 15,250,081 74.88% 5,109,358 25.12%

GROUP 07 - SUB GROUP 02


Thank
you!

GROUP 07 - SUB GROUP 02

You might also like