Underwater images are important in uses like marine research, underwater historic sites, and control of underwater drones since images are highly sensitive in such uses. Nevertheless, the use of light underwater is an issue that is limited by cases of light penetration, therefore, weak colors, contrasts, and overall vision. In order to meet these challenges, this research creates a new method transforms underwater images through Faster Region-based Convolutional Neural Network (Faster RCNN) processing in the LAB color space to deliver improved quality results. LAB color space splits light intensity channel (L) from color components (A and B channels), and there is a possibility of fine-tuning the brightness/contrast and color balance. This method helps restore natural tones and balance light very well all at once. Faster RCNN is also implemented to focus and boost important areas where maritime and underwater objects are located, and to emphasize such objects and structures while smoothing out detailed textures and secondary features. The provided method has been tested on underwater datasets illustrating higher effectiveness in comparison with common methods used in image-enhancement when such parameters as distinguishability, color correct, and object localization are expected to be improved. Peak Signal to Noise Ratio, Structural Similarity Index Measure, and color accuracy Index numerically verify the proposed approach and qualitative analysis demonstrates enhanced perception in underwater scenes. This work shows that by combining LAB based color correction with the region specific correction based on Faster RCNN, the underwater imaging issues can be resolved comprehensively and thus, this work presents a practical solution for real world applications. Not only does it enhances the appearance of the underwater images but also extends the range of their usage, thus creating further development of underwater imaging technologies.