Breaking Communication Barriers in Healthcare

Research Project · MLSP

Breaking Communication Barriers in Healthcare

A personalized Speech-to-Sign Language avatar for assistive communication

Overview

Many deaf and hard-of-hearing people struggle to understand doctors who do not know sign language, and interpreters are not always available — so vital information about symptoms, medication and treatment can be misunderstood. This project develops a Personalized Speech-to-Sign Language Avatar: the doctor speaks naturally in Sinhala and an intelligent 3D avatar translates the message into Sinhala Sign Language (SSL) for deaf patients.

Unlike systems that translate word-by-word, our approach first understands the meaning of the message and then generates natural gestures and facial expressions — making healthcare communication more accurate, understandable and human-like.

Methodology & Workflow

1

Speech Input

The doctor speaks naturally into a microphone; the system captures the spoken Sinhala speech.

2

Speech-to-Text

The recorded speech is converted into Sinhala text using speech-recognition technology.

3

Understanding the Message

The text is analysed to capture the meaning of the sentence rather than translating word by word.

4

Sign Language Translation

The sentence is mapped to Sinhala Sign Language using a sign database and gesture mappings.

5

Gesture Generation

Hand movements, body postures and facial expressions are generated to represent the message.

6

3D Avatar Animation

A 3D avatar performs the generated signs and expressions as a natural visual representation.

7

Visual Output

The sign-language animation is shown on screen so the deaf patient understands clearly.

8

System Evaluation

The system is tested and refined on translation accuracy, sign clarity and response speed.

End-to-end workflow

1

Doctor Speech

2

Speech Recognition

3

Meaning Analysis

4

Sign Translation

5

Gesture Generation

6

3D Avatar

7

Patient Understanding

Research Team

Group Member

Group Member

Undergraduate Researcher
MLSP Research Group
Group Member

Group Member

Undergraduate Researcher
MLSP Research Group