Proposes FGDPA, a reparameterizable CNN for real-time underwater image enhancement that injects DCT frequency priors and uses frequency-guided dual-path attention to reach SOTA quality at 4.23K parameters and 600+ FPS.
Aquadiff: Diffusion-based underwater image enhancement for addressing color distortion,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Real-Time Underwater Image Enhancement via Frequency-Guided Dual-Path Attention
Proposes FGDPA, a reparameterizable CNN for real-time underwater image enhancement that injects DCT frequency priors and uses frequency-guided dual-path attention to reach SOTA quality at 4.23K parameters and 600+ FPS.