# Real-time image processing

MODNet is a real-time trimap-free portrait matting model capable of efficient image segmentation and background removal.

Image segmentation model based on the Transformers library for image segmentation tasks in vision applications

A violation image classifier fine-tuned from Google's ViT model with 94.78% accuracy

A mask detection model using the ConvNeXt architecture, trained on the Face-Mask18K dataset with an accuracy rate of 99.61%.