Clinical Computer Vision on iOS: Developing BiteCheck for Insect Sting and Rash Triage
Published on 2026-10-25 · 7 min read · by Kreggscode
Clinical Computer Vision on iOS: Developing BiteCheck for Insect Sting and Rash Triage
When outdoor adventurers, parents, or travelers discover an unidentified bite or localized skin inflammation, prompt identification is vital. Differentiating a harmless mosquito welt from a dangerous brown recluse spider bite or a Lyme-carrying blacklegged tick bite can prevent serious complications.
To bring diagnostic clarity to everyday smartphones, we developed [BiteCheck AI](https://kreggscode.com/apps/bitecheck/), an intelligent computer vision application designed for instant insect bite identification, erythema progression tracking, and clinical triage.
The app is published live on the Apple App Store:
👉 [Download BiteCheck: Bug Bite Identifier on the App Store](https://apps.apple.com/app/id6810638148)
View the developer's full portfolio on the [Subodh Godi Apple Developer Page](https://apps.apple.com/developer/subodh-godi/id1831370968).
---
1. The Engineering Architecture Behind BiteCheck
Building clinical-grade vision algorithms on mobile devices requires navigating severe real-world variability: differing skin pigments, lighting gradients, and superficial similarities between different arthropod stings.
┌────────────────────────────────────────────────────────┐
│ BiteCheck Multi-Stage Vision AI │
└────────────────────────────────────────────────────────┘
│
▼
┌──────────────────────────────────┐
│ AVCaptureSession Auto-Focus │
│ (Macro Lens Distance Meter) │
└────────────────┬─────────────────┘
│ 4K Macro Frame
▼
┌──────────────────────────────────┐
│ Skin Lesion Segmentation │
│ (CoreML Depth & Edge Extraction)│
└────────────────┬─────────────────┘
│ Normalized Lesion ROI
▼
┌──────────────────────────────────┐
│ Multi-Head Ensemble Classifier │
│ (50+ Species Morphological Net) │
└────────────────┬─────────────────┘
│ Diagnostic Probability Distribution
▼
┌──────────────────────────────────┐
│ Clinical Rule Engine │
│ (Lyme Bullseye & Triage Logic) │
└──────────────────────────────────┘---
2. CoreML Ensemble Classification in Swift 6
BiteCheck uses an on-device ensemble architecture combining a mobile-optimized Vision Transformer (ViT) with a spatial convolutional network trained on over 50,000 clinically validated dermatological samples:
import CoreML
import Vision
import SwiftUI
actor BiteDiagnosticEngine {
private var classificationModel: VNCoreMLModel?
init() {
Task {
do {
let config = MLModelConfiguration()
config.computeUnits = .all // Leverages Apple Neural Engine (ANE)
let model = try BiteClassifier(configuration: config)
self.classificationModel = try VNCoreMLModel(for: model.model)
} catch {
print("Failed to initialize CoreML model: \(error)")
}
}
}
func analyzeLesion(image: CGImage) async throws -> BiteDiagnosis {
guard let model = classificationModel else {
throw DiagnosticError.engineNotReady
}
return try await withCheckedThrowingContinuation { continuation in
let request = VNCoreMLRequest(model: model) { request, error in
if let error = error {
continuation.resume(throwing: error)
return
}
guard let results = request.results as? [VNClassificationObservation],
let topResult = results.first else {
continuation.resume(throwing: DiagnosticError.noObservation)
return
}
let diagnosis = BiteDiagnosis(
species: topResult.identifier,
confidence: topResult.confidence,
isEmergency: EmergencyTriage.isHighRisk(topResult.identifier)
)
continuation.resume(returning: diagnosis)
}
request.imageCropAndScaleOption = .centerCrop
let handler = VNImageRequestHandler(cgImage: image, options: [:])
do {
try handler.perform([request])
} catch {
continuation.resume(throwing: error)
}
}
}
}
---
3. Erythema Migrans (Lyme Disease) Detection
One of BiteCheck's flagship safety tools is the Bullseye Rash Detection Pipeline. Lyme disease transmission often exhibits an Erythema Migrans concentric ring pattern.
BiteCheck calculates:
1. Centroid Puncture Localization: Isolates the central bite locus.
2. Concentric Radial Ring Gradient: Evaluates whether surrounding erythema displays an expanding boundary with central clearing.
3. Temporal Size Tracking: Users take daily photos under calibrated lighting; the app calculates diameter growth ($mm/\text{day}$) to alert users if the rash indicates active bacterial infection.
---
4. Key Takeaways and Experience BiteCheck
Download BiteCheck today to keep your family protected during outdoor travels: