Product Introduction
- Overview: The Attractiveness Test is a client-side, privacy-first web application that performs a computational aesthetic analysis of facial geometry using a pre-trained machine learning model. It falls under the category of AI-powered biometric analysis tools.
- Value: It provides users with an objective, data-driven assessment of facial proportions based on established canons like the rule of fifths and facial thirds, without compromising personal data privacy, as all processing occurs locally in the browser.
Main Features
- Client-Side Processing: The core 6MB TensorFlow.js or ONNX Runtime model loads once and executes entirely in the user's browser. No image data is transmitted to a server, ensuring complete privacy and GDPR/CCPA compliance by design.
- Comprehensive Geometric Analysis: The engine detects 478 facial landmarks and calculates 22 key ratios and angles, segmenting the analysis into three scored dimensions: Harmony (symmetry and proportion), Dimorphism (gender-typical structure), and Angularity (jawline definition).
- PSL Range Scoring: Instead of a single score, it outputs a PSL (Perceived Sexual Leagues) range (1-8). The range accounts for the unmeasurable "Misc" category (skin, hair, grooming) by providing a low-end (assuming weak misc) and high-end (assuming strong misc) estimate, offering a more nuanced and realistic result.
Problems Solved
- Challenge: Users seeking objective feedback on facial aesthetics are faced with either subjective human ratings on forums or privacy-invasive apps that upload sensitive biometric data.
- Audience: This tool serves adults (18+) interested in self-improvement, modeling, cosmetic procedure research, or personal curiosity, who prioritize data privacy and want metrics beyond subjective opinions.
- Scenario: An individual considering orthodontics or cosmetic procedures can get a baseline geometric analysis to identify specific proportional imbalances (e.g., eye spacing, jaw angle) to discuss with a professional, all without creating a permanent digital record of their inquiry.
Unique Advantages
- Strengths & Limitations:
- Strengths: Unmatched privacy (zero data transfer), immediate free access with no sign-up, transparent methodology showing all 22 measurements, and use of a peer-rated dataset (2,513 ratings on 101 faces) for calibration.
- Limitations: Requires a high-quality, front-facing, neutral-expression photo with good lighting. It cannot assess dynamic features (smile, expression), skin texture, hair, or grooming, which it explicitly states. The "Misc" factor creates a score range rather than a precise point.
- Alternatives & Differentiation:
- Vs. Social Media Filters: This tool provides diagnostic metrics and ratios, not just beautification. It's analytical, not augmentative.
- Vs. Server-Based Rating Apps: The key technical differentiator is the client-side AI execution. Alternatives like "Photofeeler" or "Qoves Studio" require server uploads. This tool's entire pipeline—from face detection with MediaPipe or a similar library to model inference—runs locally.
Frequently Asked Questions (FAQ)
- How accurate is the free AI attractiveness test? The test provides an accurate analysis of static facial geometry based on 22 measured proportions compared to a reference dataset. Its accuracy is highest for bone structure and symmetry from a compliant photo, but it cannot rate variable factors like skin, hair, or style, which is why it provides a score range.
- Is my photo stored or sent to a server? No, your photo never leaves your device. The entire analysis, including the AI face landmark detection and scoring model, runs locally in your web browser. No data is transmitted or stored on any server.
- What do the Harmony, Dimorphism, and Angularity scores mean? These are sub-scores (1-8) for specific geometric attributes: Harmony measures symmetry and classic proportions (e.g., facial thirds); Dimorphism assesses how masculine/feminine the bone structure is relative to a chosen reference; Angularity evaluates the definition of the jawline and chin from the front view.