
AI-Based Pose Comparison Mobile Application
A cross-platform mobile app that lets users compare their body posture and movements with a reference coach or trainer video using on-device AI pose detection.
An AI-powered event platform that transforms children's drawings into storybook-style artwork in seconds with instant printing, email delivery, and high-volume event support.

Our client wanted to create an engaging, technology-driven attraction for families attending the Forever Fairview Park event. The objective was to give every child the opportunity to see their imagination come to life by transforming hand-drawn artwork into professionally illustrated storybook-style characters within a matter of seconds.
The solution needed to operate continuously throughout the two-day event while handling thousands of drawings with minimal staff intervention. Since the experience would be delivered inside a live event booth, the application had to provide a simple workflow that allowed booth staff to quickly scan drawings, generate AI artwork, print or email the final keepsake, and immediately prepare the system for the next participant.
The primary goal was to create a memorable family experience while ensuring reliability, speed, and scalability during periods of high visitor traffic.
Children and families expect immediate engagement during public events. The complete experience—from scanning the drawing to displaying the final AI-generated artwork—needed to be completed in under one minute without creating long waiting lines.
The AI model needed to enhance children's artwork while maintaining a strong resemblance to the original drawing. The generated illustration had to look professionally illustrated without losing the unique characteristics of the child's imagination.
The platform needed to process between 1,000 and 1,500 drawings per day across multiple iPads simultaneously without degradation in performance or system reliability.
Live events leave little room for technical failures. The application needed to continue operating smoothly despite temporary network interruptions, API delays, or unexpected errors while providing quick recovery mechanisms for booth staff.
Event staff required an extremely intuitive application that minimized training requirements. Every interaction needed to involve only a few taps, allowing volunteers to focus on children rather than operating complex software.
Parents expected immediate access to their child's artwork through either high-quality printing or email delivery directly from the event booth.
We developed an AI-powered image generation pipeline that converts children's hand-drawn sketches into polished storybook-style illustrations using Stable Gemini Image-to-Image model through the Gemini API.
The system preserves the original character design while enhancing colors, textures, and artistic quality, producing visually appealing keepsakes that remain recognizable to the child's original drawing.


A dedicated iPad application was designed specifically for high-volume event operations.
The application guides booth staff through a streamlined six-step workflow consisting of:
Large touch controls, landscape orientation, animated transitions, and minimal user interaction allow staff to process participants quickly while maintaining an engaging experience for families.
Once a drawing is scanned, the backend automatically performs image preprocessing before sending it to the AI model.
The workflow includes:
This automated pipeline ensures every generated illustration maintains consistent quality while reducing manual intervention.


The backend was designed to support continuous image generation throughout the event.
Using scalable cloud infrastructure and asynchronous processing, the platform efficiently manages concurrent requests from multiple iPads while maintaining consistent response times.
Load testing validates that the system can comfortably support the expected event traffic without performance bottlenecks.
After the AI artwork is generated, families can immediately receive their keepsake through multiple delivery options.
Staff can:
Every generated illustration includes the Forever Fairview Park watermark while maintaining high-resolution output suitable for printing.

The platform is deployed using scalable cloud infrastructure capable of handling thousands of image generation requests during short periods of high activity.
The architecture includes:
This infrastructure enables the platform to scale beyond a single event and support future activations with minimal configuration changes.
Given the time-sensitive nature of live events, the platform includes comprehensive monitoring and operational support.
The backend continuously tracks API performance, image generation requests, processing times, and system health. Detailed request logging enables rapid diagnosis of issues, while remote deployment capabilities allow fixes to be applied immediately without interrupting event operations.
Dedicated live technical support throughout the event ensures that any critical issues can be resolved with minimal downtime.
Children instantly see their artwork transformed into professionally illustrated storybook characters, creating a unique and memorable event attraction.
The optimized workflow allows staff to process thousands of drawings over a two-day event while maintaining short wait times and smooth visitor flow.
Image preprocessing and AI optimization produce high-quality illustrations that closely resemble each child's original drawing while enhancing artistic presentation.
The intuitive iPad application enables volunteers and event staff to operate the entire experience with minimal training, reducing operational complexity.
The modular cloud architecture allows the solution to be reused for future exhibitions, shopping malls, festivals, educational events, and brand activations without requiring significant redevelopment.

Explore other projects we've delivered

A cross-platform mobile app that lets users compare their body posture and movements with a reference coach or trainer video using on-device AI pose detection.

An interactive AR educational app for children, combining 3D characters, real-time AI interactions, and book-based learning.

Transforming screen recordings into actionable performance insights with high-fidelity computer vision and neural networks for competitive gaming.