Co esponding au ho : Yogi ha Se y
Copy igh © 2025 Au ho (s) e ain he copy igh o his a icle. This a icle is published unde he e ms o he C ea i e Commons A ibu ion License 4.0.
Su ey on Deco ision : Rede ine you space wi h sma deco
Ka i ha Soppa i 1, V. Re a hi Chand ika 2, Yogi ha Se y 2, * and O.Sakshi h 2
1 Head & P o esso CSE-AI&ML Depa men , ACE Enginee ing College, Hyde abad, India.
2 S uden s o Depa men CSE (AI&ML) o ACE Enginee ing College, ACE Enginee ing College, Hyde abad, India.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 1931-1936
Publica ion his o y: Recei ed on 29Ma ch 2025; e ised on 11 May 2025; accep ed on 13 May 2025
A icle DOI: h ps://doi.o g/10.30574/wja .2025.26.2.1645
Abs ac
This su ey in oduces Deco ision, a sma and modula in e io design pla o m ha uses AI, machine lea ning, and
web echnologies o o e a pe sonalized home deco expe ience. I in eg a es eal- ime oom isualiza ion ia S abili y
AI's image gene a ion API, a s yle ecommenda ion engine based on a use quiz, and a u ni u e cha bo ha il e s IKEA
p oduc s using na u al language que ies. Buil wi h Flask and SQLi e, he pla o m emphasizes accessibili y and cos -
e iciency h ough open-sou ce ools. Deco ision’s scalable a chi ec u e suppo s a human-cen e ed design app oach,
helping use s ans o m p e e ences in o ac ionable design ou pu s ega dless o expe ise. This su ey highligh s he
sys em’s po en ial o simpli y and democ a ize in e io s yling h ough in ui i e AI-d i en in e ac ions.
Keywo ds: AI In In e io Design; S abili y AI; S yle Recommenda ion Engine; Flask; Sqli e; NLP; Use -Cen e ed Design;
Fu ni u e Cha bo ; Home Déco ; Image Gene a ion
1. In oduc ion
In e io deco a ion is a subjec i e and c ea i e p ocess ha o en lacks a o dabili y, guidance, and accessibili y o
e e yday use s. Wi h he ise o e-comme ce and AI, he e is an oppo uni y o democ a ize design by combining da a,
p e e ences, and gene a i e ools.
Eme ging AI capabili ies o e he po en ial o enhance pe sonaliza ion and simpli y decision-making in in e io design.
The inc easing a ailabili y o isual gene a ion models, s uc u ed da a sou ces like p oduc in en o ies, and accessible
web echnologies c ea es he ideal se ing o a sys em ha b idges c ea i i y wi h compu a ion. Ou mo i a ion lies in
c ea ing a cohesi e, use -cen e ed applica ion ha lowe s ba ie s o en y while p o iding high-quali y design suppo .
S abili y AI helps use s isualize and build oom aes he ics by iden i ying sui able hues om inspi a ion images, hus
aligning design sugges ions wi h hei isual p e e ences. I adds an ex a laye o pe sonaliza ion o he o e all use
expe ience.
The in eg a ion o a i icial in elligence in o in e io design enables new o ms o cus omiza ion, c ea i i y, and use
accessibili y. The shi om adi ional s a ic design ools o in elligen sys ems e lec s he g owing need o
pe sonalized and dynamic expe iences. This pape in oduces Deco ision, an AI-powe ed modula design sys em ha
b idges use ’s
s ylis ic p e e ences wi h machine-gene a ed design ou pu s. I emphasizes use empowe men by allowing e e yday
indi iduals o engage in he design p ocess h ough accessible ools like quizzes, cha bo s, and isual ans o ma ions.
The objec i e is o build a scalable and cos -e ec i e solu ion ha suppo s c ea i i y, unc ionali y, and eal-wo ld
usabili y.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 1931-1936
1932
2. Li e a u e e iew
2.1. Visually Compa ible Home Déco Recommenda ions - Zhang e al. (2022)
This s udy p esen s a sys em ha le e ages AI o ma ch u ni u e and accesso ies wi h exis ing oom aes he ics based
on isual compa ibili y and s yle p e e ences.
Me hodologies Used: Con olu ional Neu al Ne wo ks (CNNs) o isual analysis and ma ching oom aes he ics wi h
u ni u e design.
2.2. Pe sonalized In e io s a Scale - Zhou & Wang (2024)
The pape discusses scaling pe sonalized in e io design ecommenda ions by analyzing use in e ac ion and
p e e ences.
Me hodologies Used: Da a analy ics and use beha io analysis o p o ide pe sonalized ecommenda ions based on
la ge da ase s.
2.3. Home Déco Recommenda ions Using Collabo a i e Fil e ing - Choi e al. (2023)
This esea ch explo es using collabo a i e il e ing algo i hms o ecommend home déco based on use beha io and
p e e ences.
Me hodologies Used: Collabo a i e Fil e ing, which iden i ies use p e e ences and sugges s déco based on simila
beha io s om o he use s.
2.4. In e ac i e In e io Design Recommenda ion Sys em (IIDRS) - Zhang e al. (2024)
This sys em in eg a es use p e e ences wi h AI o gene a e pe sonalized oom layou s and déco sugges ions h ough
in e ac i e inpu s like oom ype and budge .
Me hodologies Used: AI-based design in eg a ion, using s uc u ed use inpu da a o gene a ing pe sonalized oom
layou s.
2.5. Deep Lea ning-Based Home Déco Recommenda ion Sys em - Li e al. (2023).
This pape p oposes a deep lea ning sys em o ecommending u ni u e and déco , u ilizing use da a and con ex -
awa e ea u es.
Me hodologies Used: Deep Lea ning, Use Beha io Da a Analy ics, Con ex -Awa e Fil e ing.
2.6. AI-D i en Pe sonalized Design o Home In e io s - Wang e al. (2023)
This esea ch ocuses on gene a ing pe sonalized home déco s yles using AI and design analysis based on use inpu
and p e e ences.
Me hodologies Used: AI algo i hms, p e e ence analysis, and neu al ne wo k-based s yle gene a ion.
2.7. Home Deco a ion wi h Gene a i e Design Models - Chen e al. (2023).
The s udy discusses how gene a i e design models can be used o assis in home deco a ion by c ea ing pe sonalized
layou s and ecommenda ions based on use speci ica ions.
Me hodologies Used: Gene a i e Design Models, Neu alNe wo ks o pe sonalized ecommenda ions, and s yle
gene a ion.
2.8. Home In e io Design Using NLP and AI - Pa el e al. (2024).
The pape in oduces a na u al language p ocessing-based in e ace o in e ac wi h use s and sugges home déco ,
ocusing on pe sonaliza ion.
Me hodologies Used: Na u al Language P ocessing (NLP), Use Que y Unde s anding, AI o Recommenda ion Sys ems.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 1931-1936
1933
2.9. Real- ime Pe sonalized Home Déco Sys em - Liu e al. (2024)
This wo k p esen s a eal- ime sys em ha o e s pe sonalized home déco ecommenda ions based on dynamic use
inpu s and design p e e ences.
Me hodologies Used: Real-Time Da a Analy ics, Dynamic Use P e e ences, AI-based Room Design Assis ance.
2.10. AI-Based In elligen Home Déco Sys em Lee e al. (2023).
This sys em ocuses on AI-d i en analysis o use ’s ooms o sugges op imal u ni u e placemen s and déco
combina ions.
Me hodologies Used: AI, Deep Lea ning Algo i hms, Room Analysis o déco sugges ions.
3. Objec i es
This p ojec aims o build an AI-powe ed in e io design pla o m ha o e s use s he abili y o upload o desc ibe hei
ooms and ecei e isually enhanced edesigns gene a ed h ough S able Di usion. I will include an in e ac i e s yle
iden i ica ion quiz ha connec s use p e e ences o speci ic in e io design s yles using a s uc u ed da ase . The
pla o m will also ex ac colo pale es om oom images by applying OpenCV and clus e ing algo i hms such as
KMeans. Pe sonalized u ni u e and deco ecommenda ions will be p o ided based on he use ’s design s yle and
budge , using NLP-d i en il e ing and eal- ime p oduc da a. All componen s including image ans o ma ion, s yle
quiz, colo pale e gene a ion, and u ni u e planning will be seamlessly in eg a ed in o a esponsi e web in e ace
de eloped wi h Reac and Flask. Use da a and quiz ou comes will be e icien ly managed using SQLi e o Pandas-based
logic o suppo ongoing pe sonaliza ion. Ul ima ely, he pla o m aims o deli e a ully au oma ed and ailo ed design
expe ience,
4. A chi ec u e
The AI-powe ed in e io design pla o m is s uc u ed o p o ide a smoo h and in elligen use expe ience by combining
use inpu , sma p ocessing, and eal- ime da a. Use s upload images o desc ibe hei ooms ia an in e ace, which
igge s he co e sys em o p ocess he inpu .
Figu e 1 Sys em a chi ec u e
4.1. Deco ision AI sys em a chi ec u e o pe sonalized home déco ecommenda ions.
The AI engine hen pe o ms asks like s yle de ec ion, oom edesign, and pe sonalized ecommenda ions, wo king
alongside he business logic ha handles design, ans o ma ion, and budge alignmen . Ex e nal APIs p o ide eal- ime
p oduc da a, while he da a s o age module e ains use inpu s and esul s o pe sonaliza ion. Final ou pu s a e
deli e ed h ough he in e ace, o e ing ailo ed and isually enhanced design sugges ions.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 1931-1936
1934
5. Expe imen al Resul s and Pe o mance Compa ison
This sec ion compa es ecen image gene a ion models, ocusing on hei eal- ime ende ing speed and key isual
ea u es like image sha pness, s ylis ic accu acy, and use app o al. The goal is o assess how well each model balances
as p ocessing wi h high isual ideli y, which is essen ial o smoo h use in e ac ions in in e io design pla o ms.
This compa ison p o ides insigh s in o he s eng hs and limi a ions o models used o AI-based oom edesign asks.
Figu e 2 Pe o mance Compa ison
5.1. Algo i hm Compa ison: Visual s. Pe sonaliza ion
The ba cha "Algo i hm Compa ison: Visual s. Pe sonaliza ion" compa es six algo i hms—CNNs, Collabo a i e
Fil e ing, Deep Lea ning, Gene a i e Models, NLP, and Da a Analy ics—on isual analysis and pe sonaliza ion (0–10
scale). Deep Lea ning sco es highes and mos balanced in bo h, making i he mos e sa ile. CNNs excel in isual
analysis bu lag in pe sonaliza ion, while Collabo a i e Fil e ing and Da a Analy ics show he opposi e end, p io i izing
pe sonaliza ion o e isuals. Gene a i e Models pe o m well in bo h, hough sligh ly below Deep Lea ning. NLP sco es
lowes o e all. The cha sugges s combining Deep Lea ning andGene a i e Models o sys ems needing bo h isual and
pe sonalized ecommenda ions, wi h Collabo a i e Fil e ing and Da a Analy ics as suppo ing pe sonaliza ion ools.
Table 1 Compa ison able
S.No
Ti le & Au ho
Focus A ea
Me hodologies
Used
Con ibu ions
Limi a ions
1
Visually
Compa ible Home
Déco
Recommenda ions
Zhang e al. (2022)
Ma ching
u ni u e &
accesso ies wi h
oom aes he ics
based on isual
compa ibili y
CNNs o isual
analysis
Achie ed s yle-
ma ching
be ween exis ing
déco and new
i ems isually
Limi ed o s a ic
images; s uggles
wi h dynamic
en i onmen s
2
Pe sonalized
In e io s a Scale
Zhou & Wang
(2024)
Scaling
pe sonalized
in e io design ia
use in e ac ion
analysis
Da a Analy ics,
Use Beha io
Analysis
Scaled
ecommenda ions
o la ge da ase s,
enabling
pe sonaliza ion a
scale
Lacks isual
analysis; mainly
beha io -d i en
wi hou aes he ic
alida ion
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 1931-1936
1935
3
Home Déco
Recommenda ions
Using
Collabo a i e
Fil e ing Choi e al.
(2023)
Recommending
déco based on
simila use
beha io
Collabo a i e
Fil e ing
Le e aged use
simila i ies o
be e
pe sonaliza ion
Cold-s a
p oblem; poo
pe o mance o
new use s
wi hou his o y
4
In e ac i e
In e io Design
Recommenda ion
Sys em
(IIDRS)Zhang e al.
(2024)
Pe sonalized
layou s & déco
ia in e ac i e
inpu s
AI-based design
in eg a ion,
S uc u ed Use
Inpu
Enabled use s o
speci y budge ,
oom ype o
ailo ed
ecommenda ions
Limi ed
au oma ion; elies
hea ily on
s uc u ed inpu
5
Deep Lea ning-
Based Home
Déco
Recommenda ion
Sys em Li e al.
(2023)
Recommenda ions
using use da a &
con ex ual ac o s
Deep Lea ning,
Use Da a
Analy ics, Con ex -
Awa e Fil e ing
Inco po a ed
con ex ual
awa eness o
mo e ele an
sugges ions
Requi es la ge
labeled da ase s;
scalabili y issues
6
AI-D i en
Pe sonalized
Design o Home
In e io s Wang e
al. (2023)
Gene a ing
pe sonalized
déco s yles wi h
AI & design
analysis
AI Algo i hms,
P e e ence
Analysis, Neu al
Ne wo ks
Combined use
p e e ences wi h
neu al s yle
gene a ion o
unique designs
Lacks eal- ime
adap abili y;
ocused on s a ic
ecommenda ions
7
Home Deco a ion
wi h Gene a i e
Design Models
Chen e al. (2023)
Pe sonalized
layou s & déco
using gene a i e
models
Gene a i e Design
Models, Neu al
Ne wo ks
Enabled
au oma ic design
gene a ion based
on use specs
May gene a e
imp ac ical o
un ealis ic
designs; needs
human alida ion
8
Home In e io
Design Using NLP
and AI Pa el e al.
(2024)
Pe sonalized
déco
ecommenda ions
ia NLP-based
in e ac ion
NLP, Use Que y
Unde s anding, AI
Recommenda ions
Allowed use s o
in e ac na u ally
ia language o
ecommenda ions
Limi ed
unde s anding o
complex que ies;
language
ambigui y issues
9
Real- ime
Pe sonalized
Home Déco
Sys em Liu e al.
(2024)
Real- ime
ecommenda ions
based on dynamic
use inpu
Real-Time Da a
Analy ics,
Dynamic
P e e ences, AI
Design
Enabled adap i e
ecommenda ions
as p e e ences
change in eal-
ime
Compu a ionally
in ensi e;
challenges in
la ency &
p ocessing
10
AI-Based
In elligen Home
Déco Sys em Lee
e al. (2023)
Sugges ing
op imal u ni u e
placemen & déco
ia oom analysis
AI, Deep Lea ning,
Room Analysis
Focused on space
op imiza ion &
unc ional
placemen
Visual ou pu may
lack aes he ic
appeal; depends
on oom pho o
quali y
6. Conclusion
The analysis o AI-powe ed home deco ecommenda ion sys em e eals signi ican ad ancemen s and pe sis en
challenges in he ield. Deco ision success ully demons a es how a i icial in elligence can e olu ionize in e io design
by making i mo e accessible, pe sonalized, and in e ac i e. By in eg a ing S abili y AI’s image gene a ion models, a
cus om s yle quiz, and eal- ime IKEA p oduc ecommenda ions, he sys em o e s a seamless design expe ience om
inspi a ion o implemen a ion. Use s can isualize oom ans o ma ions, disco e hei design p e e ences, and explo e
ma ching p oduc s wi hin hei budge . The combina ion o ex - o-image syn hesis, use -d i en inpu , and e-comme ce
in eg a ion makes DecoVision a unique and p ac ical solu ion o mode n in e io design challenges. Howe e , issues
such as da ase limi a ions, isual compa ibili y, and he need o dynamic, adap i e adjus men s o use p e e ences
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 1931-1936
1936
and en i onmen al ac o s s ill exis . To add ess hese challenges, u u e esea ch should ocus on mul imodal AI models
ha inco po a e beha io al and con ex ual da a o mo e pe sonalized ecommenda ions. By add essing hese gaps,
Deco ision and simila AI sys ems ha e he po en ial o se new s anda ds o pe sonalized, e icien , and accessible
in e io design. To imp o e and scale he DecoVision sys em u he , he ollowing enhancemen s a e p oposed.
Compliance wi h e hical s anda ds
Disclosu e o con lic o in e es
The au ho s decla e ha he e is no con lic o in e es .
Re e ences
[1] Rad o d e al., “Lea ning T ans e able Visual Models F om Na u al Language Supe ision”, a Xi :2103.00020,
2021.
[2] Ramesh e al., “Ze o-Sho Tex - o-Image Gene a ion”, a Xi :2102.12092, 2021.
[3] Saha ia e al., “Pho o ealis ic Tex - o-Image Di usion Models wi h Deep Language Unde s anding”,
a Xi :2205.11487, 2022.
[4] Chang e al., “A Deep Lea ning App oach o Realis ic Room Design Using GANs”, P oceedings o CVPR Wo kshops,
2019.
[5] Zhu e al., “Visual Lea ning and Fea u e T ans e o In e io Design Recommenda ion”, ACM T ansac ions on
G aphics (TOG), 2020.
[6] sola e al., “Image- o-Image T ansla ion wi h Condi ional Ad e sa ial Ne wo ks”, CVPR, 2017.
[7] Liu e al., “In e io Design Recommenda ion Based on Room Image Aes he ics and S yle”, ACM IUI, 2021.
[8] Wang e al., “DesignGAN: Gene a i e Ad e sa ial Ne wo k o In e io Design”, IEEE T ansac ions on Mul imedia,
2022.
[9] Zhang e al., “Pix2Pix and Beyond: Su ey o Image T ansla ion wi h GANs o Design Applica ions”,
a Xi :2303.11048, 2023.
