Core Concepts
CFAT introduces a novel triangular window technique in image super-resolution, enhancing performance and overcoming limitations of rectangular windows.
Stats
Transformer-based models have shown a 0.7 dB performance improvement over other state-of-the-art SR architectures.
The proposed model delivers superior SR results across multiple benchmark datasets.
The computational cost of dense TW-MSA and sparse TW-MSA are drastically improved as L2 ≪HW in equation 10 and ( HW S )2 ≪(HW)2 in equation 11, respectively.