You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
ARNIQA evaluates the technical quality of an image without the need for a high-quality reference and demonstrates a higher correlation to human judgments and stronger generalization capabilities than popular state-of-the-art NR-IQA metrics. ARNIQA only requires a single forward pass of a ResNet50 + linear layer for inference, and will not bring any new dependencies.
Pitch
ARNIQA module and functional support
Additional context
I've already read the contributing guidelines of torchmetrics and added ARNIQA to a popular IQA library, so I would be happy to take care of the implementation.
The text was updated successfully, but these errors were encountered:
🚀 Feature
Add a No-Reference Image Quality Assessment (NR-IQA) metric - ARNIQA (Agnolucci et al., 2024)
Motivation
ARNIQA evaluates the technical quality of an image without the need for a high-quality reference and demonstrates a higher correlation to human judgments and stronger generalization capabilities than popular state-of-the-art NR-IQA metrics. ARNIQA only requires a single forward pass of a ResNet50 + linear layer for inference, and will not bring any new dependencies.
Pitch
ARNIQA module and functional support
Additional context
I've already read the contributing guidelines of
torchmetrics
and added ARNIQA to a popular IQA library, so I would be happy to take care of the implementation.The text was updated successfully, but these errors were encountered: