Noise removal from images, or image denoising, is a critical task in computer vision, aiming to enhance image quality by suppressing unwanted random variations while preserving essential features. Convolutional Neural Networks (CNNs) have emerged as a powerful tool for this purpose, leveraging their ability to learn hierarchical representations of data. This survey provides a comparative analysis of various CNN-based techniques used for noise removal, focusing on their architectures, performance, and unique contributions. The effectiveness of denoising techniques is typically evaluated using metrics like Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM). Higher PSNR and SSIM values indicate better denoising performance. CNN-based methods, particularly those with deeper architectures, generally achieve higher PSNR and SSIM scores compared to traditional methods like Gaussian filtering or wavelet-based denoising. CNN-based techniques for noise removal have demonstrated significant advancements in both performance and versatility. Traditional deep CNNs, autoencoder-based approaches, GANs, and RNNs each offer unique strengths and trade-offs. Future research is likely to focus on improving computational efficiency, robustness to diverse noise types, and the ability to generalize across different domains, thereby broadening the applicability of CNN-based denoising methods in real-world scenarios.