FreeTransformSR: Efficient Lightweight Image Super-Resolution via Free Low-Rank Learnable Transform
FreeTransformSR is a lightweight super-resolution network that reconstructs high-resolution images from low-resolution inputs using a channel-wise free low-rank learnable transform. It achieves competitive performance with significantly fewer parameters and FLOPs than existing methods by employing adaptive feature modulation, dynamic fusion, and sample-level intensity modulation.