ImageSafetyConfig

Options & Configuration · AiDotNet.Safety

Configuration for image safety modules.

For Beginners

These settings control how images are checked for harmful content. NSFW detection catches adult content, violence detection catches graphic imagery, and deepfake detection identifies AI-manipulated faces.

How It Works

References:

  • UnsafeBench: GPT-4V achieves top F1 across 11 categories (Qu et al., 2024)
  • Vision Transformers outperform CNNs for sensitive image classification (2024)

Constructors

Constructor Summary
ImageSafetyConfig()

Properties

Property Summary
bool? DeepfakeDetection { get; set; } Gets or sets whether deepfake/AI-generated image detection is enabled.
bool? NSFWDetection { get; set; } Gets or sets whether NSFW (sexual content) detection is enabled.
bool? ViolenceDetection { get; set; } Gets or sets whether violence detection is enabled.
double? NSFWThreshold { get; set; } Gets or sets the NSFW detection threshold (0-1).
double? ViolenceThreshold { get; set; } Gets or sets the violence detection threshold (0-1).
string ClassifierType { get; set; } Gets or sets the type of image safety classifier to use.