Image enhancement is crucial for various applications like surveillance, remote sensing, and digital photography. Recent advancements in deep learning have improved accuracy, efficiency and it is being integrated in all most all sectors including Geo-spatial, Healthcare . However, existing methods often neglect the importance of individual channels and focuses on enhancing the entire image uniformly. Addressing this, we are proposing a Channel-Based Image Enhancement (CBIE) model. CBIE model aims to improve the pixel quality of the image and then recombine these enhanced channels after a Stacked DE-Processing technique the enhanced individual channels - arrangement of pixel to form pattern. Utilizing separate channel based models allows us to focus on channel-specific features instead of whole image features at once. To facilitate the assessment of image quality, we have introduced a Quality Assessment & Measurement (QAM) model. QAM model employs a K-shot ( \(k = 0-12\) ) learning technique to calculate pixel-wise Mean Square Error (P-MSE) and peak signal-to-noise ratio (PSNR) as a quantifying metric. It does so by pairing images with low, average and high quality images. Through these innovative approaches, we aim to significantly improve the spatial usability and analytical potential of images. The CBIE model achieved channel based aggregated achieves PSNR values of 41.54 (validation) and 43.85 (training) and QAM model achieved P-MSE values of 0.000041143 (validation) and 0.00002880 (training) at k = 6 on image-resolution dataset whereas for OCTA-500 CBIE model achieved PSNR of 25.83 (validation) and 27.61 (training), QAM model achieved P-MSE of 0.011 (validation) and 0.0542 (training) at k = 10.