Illuminating Fashion: Enhancing Virtual Try-On

Experiences with Adjustable Lighting in AR and

VR
Srimathy S,
3rd Year, Apparel technology
Anna University.
 INTRODUCTION
Psychological & Functional Reasons:
Social & Professional Settings:
● People choose specific colors for work, casual, and party wear
based on lighting and environment (e.g., warm colors for
evening parties, neutral tones for offices).
Cultural & Fashion Trends:
● In many cultures, color significance varies (e.g., bright colors for
celebrations, muted tones for formal events).
Matching Accessories & Makeup:
● A garment’s color under different lights affects how well it pairs
with shoes, bags, or makeup.
Online Shopping Expectations:
● Customers want the color they see on-screen to match reality,
but lighting variations cause discrepancies.
Increasing AR/VR Adoption
Rise of Virtual Try-Ons:

● Brands like Zara, Gucci, and Nike are integrating AR-based try-ons
to improve online shopping.

Metaverse & Digital Fashion:

● Virtual avatars and 3D clothing in the metaverse are shaping the

future of fashion.

● L’Oréal's AR makeup try-on and Nike's AR shoe fitting have

shown increased customer confidence in online purchases.
 PROBLEM STATEMENT
Color Inconsistency Issue: Cloth’s colour Existing AR and VR solutions don’t
appear different in various lighting factor in real-world lighting (daylight,
conditions, affecting shopping decisions. LED, fluorescent, etc.).

Need for a system

that adjusts garment
appearance based on
lighting variations.

High return rates in fashion retail and e-commerce.

● A study found that inaccurate color representation on e-commerce sites leads to customer dissatisfaction and
potential loss of sales.
● In e-commerce, color dissatisfaction is a significant driver of garment returns, with studies showing that
around 16% of online apparel purchases are returned due to color issues, often exceeding expectations or not
matching the product page.(Survey by CORESIGHT).
LIGHT TEMPERATURE TONES

COLOUR INCONSISTENCY
 EXISTING TECHNOLOGY

Virtual Fitting Rooms

Avatar-Based Virtual Fitting Room (3D AR-Based Virtual Fitting Room (Camera
Model Approach) Superposition Approach)
● preset lighting ● relies on Studio lighting
● The environment does not change dynamically
 PROPOSED SOLUTION

PHYSICALLY BASED RENDERING

( a method that simulates how light interacts with materials)

High Dynamic Range (HDR) Real-Time Rendering

environment maps which are special
Pipeline
images that store detailed lighting
information

● Multi-View Lighting ● Input Processing (Scene &

Reconstruction Camera Setup)
● Parametric Directional ● Geometry Processing
● Rasterization
Lighting Estimation
● Fragment Processing & Shading
● Output Display
 HDR Maps Dual Camera Setup

Front facing camera - captures the user and

Clothes to be tried on a different garment from the front.
environment. Back camera- captures the lighting from
different angles around the user.
PBR shader-Directional Lighting Model (to simulate how much light each fabric pixel should
receive)

Lambertian Reflection Model Specular Reflection Model

I=L⋅ρ⋅max⁡(0,N.L) Is​=L⋅ρs​⋅(N⋅H)s
Where:
Where:
● I = Final pixel intensity (how bright the fabric
● Is = Specular intensity (how strong the shiny
appears).
reflection is).
● L = Light source intensity (depends on the
● L= Light source intensity.
environment).
● ρs​= Specular reflectance (how reflective the
● ρ = Surface reflectance (depends on fabric
surface is, e.g., silk is higher than cotton).
material).
● N = Surface normal (direction the fabric is
● N = Surface normal (direction the fabric surface
facing).
is facing).
● H = Halfway vector (average of light direction L
● L= Light direction (direction of incoming light).
and view direction V:
● max(0,N⋅L) = Cosine factor, which ensures light
H=(L+V)/∣L+V∣
only affects surfaces that are facing it.
● s = Shininess factor (higher for glossy surfaces,
lower for matte).
SPECULAR
REFLECTION
Dual Camera Setup Use HDR Environment maps or Geometry Processing
real time light settings. Based on
user input.

Rasterization Fragment Processing & Shade Output Display

● HDRP (for realistic lighting & ● ●
Body Tracking (for avatar-based Light Detection
shading) fitting) (adjusting fabric shading
dynamically)
● Shader Graph (for custom ● Face Tracking (for AR overlay-
fabric materials) based try-ons) ● Color Correction
(making sure the virtual
● Cloth Physics (for garment ● Light Estimation (for realistic garment looks natural in
movement) garment adaptation to all lighting)
surroundings)
● Pose & Edge Detection
(improving garment
placement accuracy)
 VR Based Fitting Rooms
User wears the VR headset (e.g., Meta Quest, HTC Vive, etc.)

VR SDK tracks head, hand, & body movements,3D rendereing.

OpenCV analyzes lighting & adjusts garment textures accordingly.

User interacts with the clothing (change outfits, adjust fit, etc.).

3D Garment is rendered in real-time on a full-body avatar in VR.

Final scene is rendered with HDRI & global illumination for realism
 ADVANTAGES
● Enhanced User Experience – More realistic virtual try-ons.
● Reduced Returns in E-Commerce – Fewer color-related mismatches.
● Fabric-Specific Accuracy – Accounts for texture, shine, and Colour
inconsistency.
● Scalability – Can be integrated into existing e-commerce websites &
metaverse stores.
● Cross-Platform Compatibility – Works on both Android & iOS using
AR Foundation.
 FUTURE IMPLICATIONS
AI for Real-Time Adaptation – ML
algorithms detect user’s real
environment lighting and adapt
garment colors accordingly.
VR Shopping Experiences – Fully
immersive metaverse stores with hyper-
realistic lighting.
Mobile & Smart Mirror Integration –
Smart AR mirrors in retail stores.
 CONCLUSION
● Current AR try-ons lack lighting realism.

● This approach integrates dynamic lighting & CII-based fabric

adjustments.

● This can revolutionize e-commerce fashion experiences & reduce returns.

● Future advancements in AI & AR will further improve realism.

 REFERENCE
https://coresight.com/research/the-true-cost-of-apparel-returns-alarming-return-ra
tes-require-loss-minimization-solutions/
Zhao, Y., Fanello, S., & Guo, T. (2023, February). Multi-camera lighting estimation for photorealistic
front-facing mobile augmented reality. In Proceedings of the 24th International Workshop on Mobile
Computing Systems and Applications (pp. 68-73).

Brelstaff, G., & Blake, A. (1988). 17] Detecting Specular Reflections Using Lambertian
Constraints.

https://docs.unity3d.com/Packages/[email protected]/manual/shaders-i
n-universalrp.html

https://en.wikipedia.org/wiki/Physically_based_rendering
https://learn.unity.com/tutorial/creating-physically-based-materials
https://developers.google.com/ar/develop/unity-arf/getting-started-ar-foundation

https://github.com/EnoxSoftware/ARFoundationWithOpenCVForUnityExample

https://stackoverflow.com/questions/55495030/how-to-use-the-arcore-camera-ima
ge-https://www.linkedin.com/pulse/augmented-reality-try-on-clothing-app-transfor
ming-fashion-mahavia-q7ytfin-opencv-in-an-unity-android-app
Vik, M., & Vikova, M. (2000). Colour Appearance Phenomena–Metamerism. Vlakna a Textil, 7(2), 126-127.

Feng, G. (2014). Colour information in natural scenes: frequency of metamerism and colour gamut. The
University of Manchester (United Kingdom).

Chika, N. H., Emmanuel, U., Celestine, U., Dora, N., & Okereke, G. (2021). Visual Identification of
Inconsistency in Pattern. In Applications of Pattern Recognition. IntechOpen.
THANK YOU