Skin Analysis
GlamAR’s AI based Facial Skin Analysis brings skin analysis to any digital experience. Powered by deep learning and computer vision, it evaluates visible skin characteristics using a single camera capture. It detects skin concerns, including acne, wrinkles, pigmentation, and pores, and returns annotated, visual results.
The analysis runs entirely through the user’s camera, no wearables or specialized hardware required, which makes it flexible to deploy in web or mobile environments at scale.
This section explains what insights it provides, how Skin Analysis works, and how it integrates into your GlamAR applications
What Insights It Provides?
| Metric / Parameter | Tech Name | Category | Value / Range | Annotated Visualization | Description |
|---|---|---|---|---|---|
| Skin Score | total_skin_score | Overall Health | Integer (0–100) | ✅ | Aggregate health evaluation summarizing all measured concerns. Returns a summary of the skin issue and an annotated image showing all identified problems. |
| Skin Type | skin_type | Classification | Oily / Dry / Normal / Combination | ❌ | Classifies overall skin behavior based on oil and moisture balance across facial zones. |
| Skin Tone | skin_tone | Classification | Fair / Light / Medium / Olive / Tan / Deep | ❌ | Melanin-based tone classification identifying the user's natural skin shade category. |
| Skin Shade | skin_shade | Classification | Hex color code (e.g. #D2B799) | ❌ | Returns a specific hex color code representing the detected skin shade, useful for precise shade matching. |
| Skin Age | skin_age | Classification | Estimated age | ❌ | Perceived age of the skin based on its condition; the response does not enforce a fixed 18–60 range. |
| T-Zone | zones.t_zone | Zone Classification | Oily / Dry / Normal / Combination | ✅ | Independent skin type classification for the T-Zone (forehead, nose, chin) with annotated image. Available via API. |
| U-Zone | zones.u_zone | Zone Classification | Oily / Dry / Normal / Combination | ✅ | Independent skin type classification for the U-Zone (cheeks, jawline) with annotated image. Available via API. |
| Acne | acne | Key Condition | Integer (0 – 100) | ✅ | Detects whiteheads and inflamed pimples. Returns annotated image highlighting affected areas. |
| Wrinkles | wrinkles | Key Condition | Integer (0 – 100) | ✅ | Analyzes fine lines and deep wrinkles on the forehead, eye area, and folds (nasolabial, marionette lines). |
| Pores | pores | Key Condition | Integer (0 – 100) | ✅ | Measures pore visibility and density across the forehead, cheeks, and nose. |
| Eye Bags | eye_bags | Eye Area | Integer (0 – 100) | ✅ | Measures lower eyelid puffiness and topography changes in the under-eye area. |
| Dark Circles | dark_circles | Eye Area | Integer (0 – 100) | ✅ | Detects blue (vascular) or pigmented (brown) darkening around the eyes. |
| Post Acne Scars | post_acne_scars | Specific Issue | Integer (0 – 100) | ✅ | Evaluates visibility and depth of post-acne scar tissue on the face. |
| Whiteheads | whiteheads | Specific Issue | Integer (0 – 100) | ✅ | Detects density of closed comedones (whiteheads) across the face. |
| Pigmentation | pigmentation | Specific Issue | Integer (0 – 100) | ✅ | Analyzes melanin distribution and focal pigmentation patches on the forehead, cheeks, nose, and chin. |
| Hydration | hydration | Specific Issue | Integer (0 – 100) | ✅ | Evaluates visual signs of skin moisture levels, detecting dryness or dehydration. |
| Future skin simulation | — | Projection / Forecast | N/A | ✅ | AI-based projection of how the user's skin may change over time. Use the stages returned by the configured simulation; the current API example includes an initial image, 3-Month Progress, and 6-Month Result. |
Available results depend on the analysis version and the response for that scan. The current Console has nine visible-concern controls; Redness is not one of them. Zone-specific fields are available through the API rather than as visible-concern checkboxes. Simulation images are projections, not measured follow-up scans.
Future skin simulation, Product recommendation, SkinGPT, and Product routine are separate report features. Recommendations and assistant experiences also require their catalog or assistant setup.
Score Severity Guide
Concern scores and the overall skin score use a scale from 0 to 100. For responses that specify the following three-band ranges, interpret them as shown:
| Score Range | Severity | Meaning |
|---|---|---|
| 0 – 79 | Needs Attention | The concern is prominent and may require targeted skincare |
| 80 – 95 | Average | Moderate condition with room for improvement |
| 96 – 100 | Good | Healthy range with minimal or no concern detected |
Note:
Other analysis versions can return different ranges or grading formats. Read the response's score_range, severity, and other grading fields instead of hard-coding these ranges for every response. Classification metrics (Skin Type, Skin Tone, Skin Shade, Skin Age) use categorical or numeric values instead.
How Does It Work?
GlamAR’s AI Facial Skin Analysis follows a streamlined process designed for speed, accuracy, and easy integration.
- Image Capture: The system accepts input from a live camera feed or an uploaded image.
- ML Inference: The captured face is analyzed using GlamAR’s trained Convolutional Neural Network (CNN) models.
- Result Generation: The model returns the available skin metrics, generating both structured data and annotated image outputs.
- Visual Feedback: The results are returned to the client as a JSON response, including all detected metrics and the processed image.
Capture, analysis, and report loading are separate stages. Their timing depends on the device, network, image, and selected experience.
Integration Options
Choose an integration method based on your workflow and environment:
- SDKs for camera capture and the analysis experience. Available for Web, Android, and iOS.
- APIs for backend or batch processing. Send image inputs and receive JSON results with detected metrics and annotated visuals.
- Plugins for quick, low-code setup through the supported integrations, including Shopify and Webflow.
- Console configures apps, manages subjects and reports, maintains products and routines, and provides analytics.
Note:
- Use SDKs for real-time camera mode.
- Use APIs for batch or backend-triggered scans.
Setup Workflow
A typical Skin Analysis setup follows this flow:
- In the Fynd Console, open AI facial skin analysis and create an app in Skin GPT or SDK mode. Each app generates a unique Application ID for integration and analytics tracking.
- Set up your environment (development or production), generate API keys or tokens, and configure domain whitelisting.
- Use the SDK for camera capture and the report experience, or the API for backend or batch image analysis. Configure the app's instructions, appearance, report features, and any catalog, assistant, or routines required by the experience.
- Publish configuration changes, check the app's Active status, and preview the saved app. Run sample scans to verify detection accuracy and JSON response structure, including annotated image results.
- Integrate the workflow into your live app, kiosk, or platform. Track scan volume, performance, and diagnostic trends through the Analytics dashboard.