TTTIR: Unlocking Instance-Specific State Evolution via Test-Time Training for Image Restoration
This work designs Progressive State Sequence Generation (PSSG) to construct complementary spatial-frequency target states (defining what to recover), and State Transition Evolution (STE) to adapt lightweight transition operators via a restoration-oriented TTT inner loop (determining how the features should evolve).
Kai-Han Zheng, Jun Li, Hang Guo et al.
· 0 citations