I

Is This Furry

Developed by SaintGermain
A visual classification model for detecting furry-style images (mostly personal avatars), capable of identifying various types of furry images
Downloads 32
Release Time : 1/29/2023

Model Overview

This model is primarily designed to detect furry-style images (mostly personal avatars), but theoretically it can recognize various types of furry images. Training data was collected from avatar images scraped from the Mastodon platform, with non-avatar images added to balance the dataset.

Model Features

High accuracy
Validation accuracy reached 93.3%, with F1 score of 0.938 and AUC value of 0.975, demonstrating excellent performance
Balanced dataset
Training data includes 551 furry images and 641 non-furry images to avoid misjudgment
Multi-category recognition
Can distinguish between furry avatars, regular animal avatars, kemonomimi avatars, and human avatars

Model Capabilities

Furry-style image detection
Avatar classification
Binary image classification

Use Cases

Social media analysis
Furry community analysis
Analyze the usage of furry-style avatars on social media
Content moderation
Avatar content classification
Automatically identify whether user avatars are furry-style
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase