UFMFormer: a hybrid transformer-Mamba network for underwater image enhancement via FFT

Key Points

• Underwater image enhancement is significantly improved by the hybrid transformer-Mamba network, highlighting its potential for clearer visuals.
• The enhancement method utilizes fast Fourier transform (FFT) to process images effectively while maintaining details in challenging conditions.
• Observational analysis employing a hybrid model combines traditional methods with advanced algorithms for optimal performance.
• This algorithm may enable better applications in marine biology and underwater photography, enhancing visual data collection.