What is it about?
The paper proposes a synthetic data approach that can enhance the accuracy of the model on low-resolution images. Models trained on high-resolution data perform very poorly on resolution images. To enhance the discriminant power, the model is trained on the synthetic data that helps the model to improve the identification accuracy on low-resolution images.
Featured Image
Photo by Mika Baumeister on Unsplash
Why is it important?
Generating the synthetic data simulates the real-world degradations and fulfils the need for resolution augmentation required for training the model.
Perspectives
The paper shows that the model performance can be enhanced by training the model on the resolution augmented data.
Research Scholar Ravindra Kumar Soni
Malviya national institute of technology Jaipur
Read the Original
This page is a summary of: Synthetic Data Approach for Unconstrained Low-Resolution Face Recognition in Surveillance Applications✱, December 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3571600.3571661.
You can read the full text:
Contributors
The following have contributed to this page
