MKYucel/hybrid_augment

[ICCV 2023] HybridAugment++: Unified Frequency Spectra Perturbations for Model Robustness

22
/ 100
Experimental

This project offers methods for training image recognition models to be more robust. It takes your existing image datasets and applies specialized frequency-based modifications during training. The outcome is a more reliable image classification model that performs well even when encountering slightly altered or corrupted images. This is for machine learning engineers and researchers who train computer vision models.

No commits in the last 6 months.

Use this if you need to train image recognition models that maintain high accuracy on standard images while also being resilient to common real-world corruptions, noise, or minor adversarial attacks.

Not ideal if you are looking for a pre-trained, plug-and-play image recognition solution without needing to train custom models or if you are not working with image-based deep learning.

image-recognition computer-vision model-robustness machine-learning-engineering deep-learning-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 0 / 25

How are scores calculated?

Stars

17

Forks

Language

Python

License

MIT

Last pushed

Sep 28, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/MKYucel/hybrid_augment"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.