Apple Swift Student Challenge 2026 Winner

2026-04-15
Copied!

Winner Certificate

Overview

I was selected as a Winner of Apple's programming contest, Swift Student Challenge 2026.

During a vacation in Miami, I was convinced the moment the idea came to me that it would win, and I started coding a bronchoscopy simulation app right on the beach. I've been working with Swift since my third year of university, building iOS/macOS apps, and this was one of the titles I'd always wanted to earn as an Apple user.

Coding during development

Since I built it in the spare moments of my busy daily life, I was coding everywhere—from the hair salon to even the bathroom lol

Coding during development 2 Coding during development 3

App Overview

App Name BronchoQuest
Platform iPadOS
Description An app that teaches the structure of the bronchi, the role/advantages/disadvantages of bronchoscopy, and lets you simulate bronchoscopy in a game-like experience

Technical Features

Anatomical accuracy was a non-negotiable element for a medical education app. To achieve this, we constructed the airway model from actual CT image data.

Data Source: LIDC-IDRI dataset (DICOM files)

Segmentation: Used 3D Slicer to isolate the bronchial tree from surrounding lung parenchyma and exported the mesh in OBJ format

Mesh Optimization: Removed non-manifold edges in Blender, recalculated normals for the luminal mucosa, and adjusted scale to match SceneKit's coordinate system

Format Conversion: Converted to USDZ format using Reality Converter for native integration with the Apple platform

Collision Detection: Signed Distance Field (SDF)

The greatest technical challenge was keeping the virtual bronchoscope strictly within the airway interior.

A Signed Distance Field (SDF) was generated using a Python script in Blender. A 2563 voxel grid stores the signed distance from each cell to the nearest airway wall.

  • Negative values → Inside the airway
  • Positive values → Outside the airway
  • Zero → Wall surface

At runtime, trilinear interpolation of the 8 neighboring voxels provides sub-voxel precision distance values, enabling real-time collision detection at 60 FPS. When the camera approaches a wall, the SDF gradient vector pushes it back; if completely blocked, movement is projected along the wall surface, replicating the behavior of a real bronchoscope sliding along tissue surfaces.

Accessibility

The app complies with the standards set in the accessibility sections of WCAG (Web Content Accessibility Guidelines) and Apple HIG (Human Interface Guidelines).

Additionally, it supports the following 5 languages:

🇬🇧 English 🇯🇵 日本語 🇨🇳 中文 🇫🇷 Français 🇪🇸 Español

App Screenshots

BronchoQuest screenshot 1
BronchoQuest — App Screen 1
BronchoQuest screenshot 2
BronchoQuest — App Screen 2

Demo Video

App Website

Learn more about BronchoQuest

https://bronchoquest.koseiw.com →

Contact

For inquiries about this app, including coverage or implementation requests, please reach out to:

BronchoQuest@koseiw.com