Resonant Echoes

"Bridging Deaf Children to Quality Education Recovery"

Overview

“Bridging Deaf Children to Quality Education Recovery”

Resonant Echoes is an innovative augmented reality interface that helps teachers understand the intentions of language-impaired kids, facilitating better rehabilitation and social engagement for deaf students.

Deaf children often lack sufficient language skills during rehabilitation, which hinders their socialization due to difficulties in communication. The ‘Resonant Echoes’ helps therapists overcome understanding difficulties with an XR display and embedded AI, focusing on a role-play game scenario.

Role

Solo Project, Graduation Project

Duration

2023 Aug. – 2024 (5 months) 

Director

Yingqing Xu, Head of Future Lab, Tsinghua

Category

Accessibility, Deaf education, XR, Headwear

Problem

In 2023, I volunteered at the China Deaf Center. There I encountered congenitally deaf children wearing cochlear implants and receiving speech training, which was a difficult process because of poor communication between deaf kids and therapists

Around 466M people globally suffer from hearing loss. WHO research indicates that a significant portion of these individuals could see notable improvements through hearing aids and proper rehabilitation treatments. Among therapists, 85% of my interviewees ranked ‘difficulty in understanding’ as the primary challenge.

According to WHO’s World Report on Hearing 2023, around 466 million people worldwide currently experience disabling hearing loss, expected to increase to 700 million by 2050. Without intervention, this can negatively impact various life areas.

Technologies like cochlear implants can significantly help children with severe hearing loss, particularly when paired with proper rehabilitation. Post-implantation, extensive rehabilitation and support services are crucial, with cognitive training playing a significant role.

An interview-based study was initiated at the China Deaf Rehabilitation Research Center, involving 14 rehabilitation therapists, conducted through face-to-face and telephone interviews.

Key findings:

1. 85% of interviewed therapists rank ‘difficulty in understanding what the children mean’ as the top challenge in rehabilitation teaching, far surpassing the second-ranked issue of ‘lack of attention.’ 

2. 90% of interviewees believe that difficulties in understanding significantly reduce the effectiveness of cognitive training. The primary reason is that challenges in sentence comprehension hinder the continuation of conversations.

3. 90% of interviewees expressed concern about the social integration of deaf children, highlighting it as the primary issue raised.

User Research : Shadowing

This picture shows the shadowing process of user research.

In order to grasp how therapists and deaf children communicate, their key challenges, and current solutions, I undertook both an expert interview survey and a shadowing investigation at China Deaf.

Among surveyed individuals, over 90% expressed concern about the social engagement of deaf children. To foster more interaction, they frequently resort to rephrasing or clarifying referents. While sometimes effective, these methods often fail due to difficulties in understanding sentences. Hence, difficulties in understanding children’s sentences are considered the biggest obstacle in cognitive training.

At the China Deaf’s affiliated preschool rehabilitation center, I carried out a shadowing research observing classes in both the higher and lower age brackets for a cumulative 30 hours. The main observations included language training, games, creative tasks and daily activities. 

Key findings are listed below:

1. Therapists often use paraphrasing to increase responses to deaf children (i.e., restating the child’s intent). While effective, it is sometimes disrupted by a lack of understanding of the spoken content.

2. Another method to enhance responses is to ask the child to identify the object they are referring to and repeat their intent. This approach improves mutual understanding success but is more time-consuming for therapists.

3. Effective paraphrasing or responses can greatly enhance the sense of social participation for deaf children. Otherwise, conversations become too brief, reducing kids’ enthusiasm to communicate.

Prior Arts

Opportunity: Most deaf edu products focus on providing children with linguistic training / solo gaming experiences. Is there a high-tech way that allows more kids and a teacher to participate in a game together, achieving good mutual comprehension?  

Inspiration :

Comprehensive Multi-player Gaming

Inspiration: role-play games and storytelling events are common activities at kindergartens, where it provides intense interactions between multiple players.

HMW: How might I create an AR interface that supports mutual understanding in role-play games for deaf kids and therapists?

Concept :

AR Headwear for visualizing deaf kids' intentions in roleplay games

Augmented Reality Interface
System Sketch
Framework

Experiment

In order to create an AR interface that helps with understanding, I conducted 2 tech experiments.

Experiment 1 was about sharing kids’ interested points in sights with teachers(therapists);

Experiment 2 aims to restore incomplete or distorted expressions to their intended meanings.

In the above experiments, computer vision, eye-tracking technology, and LLM were utilized.

Experiment1: Therapists can see what the child is interested in
Experimen2: Therapists can know what the child is trying to say

If you wish to get access to the source code, please see: Resonant Echoes at Github.

In this experiment, I used a large model named G-VOILA. It’s an enhanced version of the Kosmos model, integrated with the ability to read eye movement data. This model can complement unspoken language information through eye-tracking, aligning well with my concept. 

I am also one of the initiators of this technology. If you wish to learn more details about the customed LLM model, please see: G-VOILA: Can AI understand us better if eye data is shared? (research under review.)

 

Prototype

Product Design

Interface Design

Special Thanks to:

Prof. Yingqing Xu

Head of the Future Lab 
Tsinghua University

Prof. Yuntao Wang

Faculty of the PI HCI Lab
Tsinghua University

Yan Qin

Faculty of Rehabilitation Center
China Deaf

Xueqiao Wang

for 3D Printing & Design Critics

Nolibox (confidential)

#Algorithm #AI Design #Knowledge Map

NeRF Virtual Avatar Reconsturction

Utilizing NeRF technology, I achieved the reconstruction of virtual avatars from 2D photos to 3D models. A 3D model was rebuilt from 24 high-definition photographs.

Learn more about this program

SpaceFlow

An astronaut-specific tea kettle leveraging space’s zero-gravity. It’s engineered to avoid spills and enables convenient swapping of tea bags.

Tools: Rhino, Realflow

Floral Echos

An artistic phone sound amplifier with a special chamber design for audio enhancement, doubling as a vase for fresh flowers, offering both visual and auditory delight.

Tools: Rhino, Keyshot

Bugoides

A biomimetic animal based on a four-bar linkage, replicating an insect’s gait, composed entirely of one type of part.

Tools: Solidworks, Mechanical Manufacturing

→ Learn more about this program

Kinetic Art

A dynamic art installation, emulating Alexandar’s style, depicting a Chinese verse:

“You stand on the bridge, beholding the view / While from the tower, the viewers behold you / The bright moon adorns your window’s frame / And in others’ dreams, you do the same.”

i: Companion

A personal companion app that establishes a unique personality, founded on the ‘I as my own partner’ principle, by constructing a virtual self from user’s linguistic inputs.

Tool:  Swift

Moonster Music

‘Moonster’ has identified users’ tendencies to either venture beyond or stick to their comfort zones in music.

This platform allows you to explore what songs others are enjoying and adapt your monster persona to mingle in various music-listening groups.

Tools: Figma

→ Visit my video to read more

Space Article Builder

An exploration into a new reading medium. I’ve envisioned a virtual space where users can create virtual buildings, embedding articles, videos, and other information within, offering a non-linear reading experience.

Tool: Minecraft

→ Visit my video to read more

New Wave

‘New Wave’ is a memorable art performance event. It took five months of preparation and was presented to over 1,000 spectators. Yuchen served as the stage director, and the event featured a ‘vintage radio’ artistic concept, linking all performances in a time-traveling radio story.

ASL Sign Search (not released)

#UX #CV #Assistive Technology

Connect with me

yuchenyao_thu@163.com

→ LinkedIn

→ Instagram

Tabc T

TabcT (Thermal Activated Bacteria Cancer Therapy) is a temperature-controlled bacterial therapy developed by Tsinghua 2023, targeting breast cancer, and it offers the advantages of being 5C: cheap, convenient, controlled, continuous, and comprehensive. This is a collaborative project, Yuchen partook in this program as a design advisor and software developer.

→ Visit our wiki to see more

Fufu Companion

“Record and soothe depression with a soft robot”

Fufu’s diary is program aimed at helping the depression patients, especially in the situation thatCOVID-19 pandemic is separating people physically, making it harder to get offline treatment.

Fufu contains both Online And Offline (OAO) product, all the way to an integrated service system. Here we present fufu’s diary, in a very story-telling way.

→ See more details

MuSee

“Reimagine the unheard miracles in an alternate expression”

Imagine a world where no mellifluous birdsong, soft-spoken words, or majestic melodies exist. How poignant to miss these sounds! Yet, 448 million globally endure this quietude due to auditory impairment.

In 2021, a group of Tsinghua University students aspired to bridge this void. They sought to ‘play’ a symphony in lights, enabling those with hearing loss to partake in the symphony’s wonder, in a one-to-one sensory dance.

→ See more details

Unlock Fear

Current Status: IMWUT 24′ under review

“Unlock the secret of fear”

Acrophobia, the fear of heights, significantly affects overall quality of life. People who suffer from acrophobia know this fear is excessive and unreasonable, yet 6.4% of total population suffer from this issue.

We investigated the feasibility of passive sensing data for fear measurement and discern the most influential physiological and behavioral indicators. 

→ See more details

Resonant Echoes​

“Bridging Deaf Children to Quality Education Recovery”

Resonant Echoes  is an innovative augmented reality interface that helps teachers understand the intentions of language-impaired kids, facilitating better rehabilitation and social engagement for deaf students.

Deaf children often lack sufficient language skills during rehabilitation, which hinders their socialization due to difficulties in communication. The ‘Resonant Echoes’ helps therapists overcome understanding difficulties with an XR display and embedded AI, focusing on a role-play game scenario.

→ See more details

G-Voila

Current Status: IMWUT 24′ under review

“Can AI understand us better if eye data is shared?”

As developers of general artificial intelligence increasingly shift their focus towards incorporating more modalities of information such as images and sounds, we can’t help but wonder: If we provide such rich eye movement information to artificial intelligence, could it enable them to better understand us?

→ See more details

Kosmos-2 Interface (not released)

#UX #LLM #Interface

Time of City

An artistic installation that embodies a variety of light and shadow patterns within its Sisyphean cycle.

Tools: Raspberry Pi, Smart Car

→ Visit my video to read more