Attractive Faces Celebs Detection
A Vision Transformer-based facial attractiveness scoring model that returns a score between 0 and 1 for input facial images, with an accuracy of approximately 83%.
Downloads 2,473
Release Time : 8/31/2023
Model Overview
This model analyzes facial images and returns a celebrity-style attractiveness score (between 0 and 1), suitable for facial aesthetic evaluation applications.
Model Features
High Accuracy
Achieves 83.78% accuracy on the test set, balancing precision and recall.
Celebrity-Style Evaluation
Specifically evaluates facial attractiveness in a celebrity style.
ViT-Based Architecture
Utilizes the Vision Transformer base model with powerful image feature extraction capabilities.
Model Capabilities
Facial Image Analysis
Attractiveness Scoring
Binary Classification
Use Cases
Aesthetic Evaluation
Celebrity Attractiveness Scoring
Evaluates whether a facial image has celebrity-level attractiveness.
Accuracy 83.78%, F1 score 0.8378
Entertainment Applications
Photo Enhancement Suggestions
Provides attractiveness evaluation references for photo enhancement applications.
Featured Recommended AI Models