Skip to main content

System Requirements

GlamAR’s AI Facial Skin Analysis SDK delivers real-time skin diagnostics using camera-based input. To achieve high diagnostic accuracy, stable performance, and user trust, ensure your integration meets the device, camera, lighting, and network conditions outlined below.

Supported Platforms

Android SDK
Runs on Android 7.0+ (API 24) with Android 10 or higher recommended for consistent camera quality and neural model performance.

iOS SDK
Supports iOS 13+ with native-only integration.
Web-based implementations are not advised due to lower color fidelity and inconsistent exposure handling in browsers.

React Native SDK
Available for Android and iOS, leveraging native camera modules for real-time frame capture and AI inference.

Flutter SDK
Runs on Android and iOS using InAppWebView for camera access and image stream processing.

Web SDK
Supported for proof-of-concept (POC) use only.
Browser-based camera access often provides insufficient color accuracy and exposure control for reliable skin analysis.

Smart Mirrors
Compatible with Android-powered mirror devices equipped with external USB or high-resolution integrated cameras.
Ideal for in-store diagnostic experiences or professional use cases.

Important:
The Web SDK is not recommended for live skin analysis due to variable image quality across browsers and devices.
For production-grade integrations, use the native Android, iOS, or Smart Mirror SDKs.

Device Compatibility

The GlamAR Skin Analysis SDK runs efficiently on a range of Android and iOS tablets as well as smart mirror systems.
Accurate skin scanning depends on high-quality cameras, stable mounting, and proper device positioning.

Tablet Support​

Android Tablets
Android Tablets are fully supported. Works seamlessly with both the native Android SDK and the Web SDK (via Chrome).

Ideal for portable skin diagnostic setups or in-store mobile kiosks.

iPadOS
Limited support. Operates only through Safari’s camera APIs, which restrict control over exposure and white balance.

Use for demonstrations or non-diagnostic previews rather than clinical-grade analysis.

Recommended Models

  • iPad Pro (2021 +): 12 MP front camera with Apple Neural Engine; excellent for high-resolution scans.
  • Samsung Galaxy Tab S8 / S9: Powerful processors and front-facing cameras suitable for production use.
  • Lenovo P12 Pro / Yoga Tab: Mid-range Android options offering adequate optics and performance.

Tip: Mount the device in landscape orientation and ensure the user’s face is centered and evenly lit for best detection accuracy.

Smart Mirror Support​

The SDK also supports Android-powered smart mirrors and kiosk installations for professional or retail environments.

Recommended Specifications

  • Camera: Logitech Brio 4K or Intel RealSense D415
  • CPU: Intel i5 / i7 (10th Gen or newer) or Apple M1 / M2 chip
  • RAM: Minimum 8 GB
  • GPU: Not required unless performing on-device machine-learning inference
  • Display: ≥ 15 in Full HD, preferably touch-enabled

Note: Our Android SDK runs reliably on most kiosk builds. Contact GlamAR support for detailed mirror integration guidance.

Environment and Lighting

Accurate skin diagnostics depend heavily on environmental conditions. Controlled lighting, consistent camera distance, and minimal interference from reflections or shadows ensure precise color and texture analysis.

Environment Guidelines

  • Indoor Use Only: Operate in a controlled indoor setting; avoid direct sunlight and window glare.
  • Subject Distance: Maintain 12 - 18 inches (30 - 45 cm) between the camera and subject’s face for proper framing.
  • Hair & Face Preparation: Tie hair back, remove makeup, and remove glasses or accessories that may obscure facial features.
  • Face Framing: Keep the head upright, eyes open, and mouth relaxed. The entire face should remain within the camera frame.

Note: Following these environmental controls for accurate model alignment and reduces false positives during skin detection.

Lighting Conditions

  • Intensity: Maintain 500 - 1000 lux of evenly distributed ambient light.
  • Light Type: Use white LED lighting between 4000 K and 5000 K color temperature for natural skin tone rendering.
  • Avoid: Backlighting, overhead glare, or tinted light sources that distort skin color balance.
  • Recommended Setup: Place a ring light or diffused LED panel at roughly a 45° angle in front of the subject to ensure even illumination across the face.

Note: Consistent, neutral lighting enhances the accuracy of redness, pigmentation, and under-eye texture analysis.

Camera

High-quality imaging is essential for accurate skin analysis. The GlamAR Skin Analysis SDK relies on fine image detail and natural color representation to detect pores, texture, acne, and pigmentation accurately.

Camera Requirements:

Resolution

  • Minimum 8 MP, recommended 12 MP or higher for accurate skin texture detection.

Focus

  • Auto-focus is required. Avoid fixed-focus webcams or devices with aggressive beauty processing.

Frame Rate

  • Minimum 30 FPS to maintain smooth real-time capture and prevent motion blur during analysis.

Color Reproduction

  • Maintain accurate white balance and disable beauty filters or camera effects. Filters can obscure natural skin texture and compromise diagnostic accuracy.

Sensor Type

  • Use CMOS or higher-end sensors for better color depth and low-light sensitivity.

Note: Fine-grained skin features such as pores, wrinkles, and blemishes require high-resolution, true-color input. Low-resolution or filtered images significantly reduce detection accuracy.

Camera Permissions:

Android SDK

  • Requires android.permission.CAMERA declared in AndroidManifest.xml and granted at runtime.

iOS SDK

  • Add NSCameraUsageDescription to your Info.plist.
    iOS will display this message when requesting access to the device camera.

React Native

Flutter

  • Use the permission_handler plugin to request camera access dynamically.

Web SDK

  • Relies on the standard browser camera access prompt via getUserMedia() API.

Best Practice: Always ask for user consent before activating the camera. iOS specifically requires a camera usage description in app metadata to avoid runtime permission errors.

Network Requirements

The GlamAR Skin Analysis SDK relies on stable network connectivity to transmit camera input and receive AI inference results in real time.

While enterprise setups can operate partially offline through edge caching, most integrations benefit from an active, low-latency connection.

Connection Guidelines:

Upload Speed

  • Minimum: ≥ 2 Mbps
  • Recommended: ≥ 5 Mbps
  • Higher upload speeds reduce image transmission delay and improve responsiveness for real-time skin diagnostics.

Latency

  • Minimum: < 300 ms
  • Recommended: < 150 ms
  • Low latency ensures faster round-trip times between capture and analysis results.

Offline Support

  • Optional: available for enterprise smart mirrors via local inference or edge caching. Contact GlamAR support for custom deployment options.

Data Transmission Metrics:

Typical Capture -> Result Time

  • ~1.5 - 3 seconds on mobile devices
  • ~800 ms - 1.5 seconds via API or edge processing

Image Format

  • Uses JPEG or PNG, typically base64 encoded when sent through the API.

Annotated Output

  • Returns one base64 image per detected condition, alongside the original capture image. Each annotated frame highlights specific areas such as pigmentation, pores, or wrinkles.

Tip: For real-time results in client applications, use asynchronous uploads and prefetch user session tokens before capture. This prevents UI blocking and keeps the user experience smooth even under variable network conditions.

Summary

GlamAR’s AI Facial Skin Analysis SDK combines computer vision and deep learning to deliver accurate, real-time skin diagnostics across mobile, web, and smart mirror environments.

To ensure precision, maintain user trust, and deliver consistent performance, integrations must meet the recommended hardware, camera, lighting, and network conditions.

Key Takeaways:

  • Hardware: Use Android 10+/iOS 13+ devices with at least 4 GB RAM and modern CPU/GPU architecture.
  • Camera: Minimum 8 MP, auto-focus capable, with accurate color reproduction and no beauty filters.
  • Lighting: Maintain 500 - 1000 lux, neutral white LED lighting (4000 - 5000 K), evenly illuminating the subject’s face.
  • Environment: Indoor setup, stable mounting, subject positioned 12 - 18 inches (30 - 45 cm) from the camera.
  • Network: ≥ 5 Mbps upload speed, < 150 ms latency for near real-time results; edge caching available for enterprise mirrors.
  • Model Input: High-resolution JPEG/PNG (base64-encoded) for API inference.
  • Output: Annotated results per detected skin condition, plus the original capture image.