With the explosion in the number of selfies taken daily, the demand for more accurate and visually pleasing images is increasing. However, even images captured by modern cameras are prone to degradation by noise, blur, or loss of color. Although tools like Photoshop can aid in restoring such images, they require significant graphic editing skills and manual interaction. This study proposes an facial image enhancement pipeline to address these challenges through deep learning methods. We explore colorization methods in various color spaces along with denoising and deblurring techniques on images with Gaussian noise and blur. The CNN networks achieve high image quality (PSNR and SSIM) and are paired with appropriate metrics to evaluate colorfulness, noise, and focus measures. All models are combined into a single pipeline, minimizing manual interaction for image enhancement. This creates a seamless experience for users to improve the visual appeal of their photographs.