Text this: OPTNet: Optimized Pixel-Transformer Model for Adaptive Retinal Fundus Image Enhancement