Cross-Platform AR Design Preview

A mobile app for previewing designs on the body in real time before making a permanent choice.

Challenge

Previewing a design before a permanent procedure required more than placing an image on a photo; it needed to follow body movement naturally. Because iOS and Android expose different AR capabilities, stable tracking was also required on Android devices without depth sensors.

Solution

Combined ARCore SLAM, feature tracking, plane detection, and body landmarks on Android to correct position, rotation, and scale. Reduced jitter with temporal smoothing and separated React Native UI responsibilities from native AR processing.

Result

Delivered real-time AR previews on iOS and Android, including devices without depth sensors, with designs that remain naturally attached to body regions.

Team

2 members, 4 months

Technical leadership, Android AR, and API design

Role

Led Android AR architecture and cross-platform technical design.

Designed tracking strategy, native bridges, REST API specifications, and the responsibility boundaries across mobile and backend teams.

Tech Stack

FrontendReact Native / TypeScript
NativeJava / Kotlin
ARARCore / SLAM / Landmark Tracking
RenderingOpenGL ES
BackendPython / REST API / Swagger

Key Features

01

Real-time body-tracked AR previews

02

Stable tracking on Android devices without depth sensors

03

Design and metadata delivery through a REST API

04

React Native bridge to Java and Kotlin AR modules

Technical Highlights

Platform-Specific AR Strategy

Used each platform's strongest AR capabilities instead of forcing identical implementations across iOS and Android.

Tracking Quality on Mid-Range Devices

Combined body landmarks with temporal smoothing to reduce jitter and preserve the appearance of a design attached to the skin.