This repository contains the implementation of a pipeline for face detection and image deblurring using advanced machine learning models. Face detection uses RetinaFace, and deblurring is performed using the GAN Prior Embedded Network (GPEN).
- RetinaFace: A state-of-the-art face detection model that excels in detecting faces within various environments and conditions.
- GPEN: Utilizes Generative Adversarial Networks (GANs) to deblur images, effectively restoring details with high precision.
The models are adapted from the GPEN repository.
- Face Detection and Deblurring Logic: Code to extract and process frames with detected faces and apply deblurring techniques.
- Sample Outputs: A folder containing before-and-after images showcasing the deblurring results.
To use this repository for your face detection and deblurring projects, follow these steps:
- Visit the Kaggle notebook at Noise Removal GANs Pipeline.
- Copy the notebook to your own Kaggle account.
- Adapt and run the code according to your specific requirements.
For collaboration inquiries, contact via LinkedIn.
Licensed under the MIT License.