Advancements in image quality assessment: a comparative study of image processing and deep learning techniques

Abstract

Image quality assessment (IQA) is a crucial field in image processing that ensures optimal performance in various applications such as medical imaging, surveillance, and multimedia systems. The evolution of IQA methods spans from traditional image processing techniques to the incorporation of advanced deep learning algorithms. This literature review aims to provide a comprehensive analysis of the methodologies used in image quality assessment, focusing on both full-reference, reduced-reference, and no-reference approaches. Traditional methods such as peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) are discussed alongside more recent deep learning-based approaches that leverage convolutional neural networks (CNNs), generative adversarial networks (GANs), and transformers for feature extraction and prediction. Deep learning models have demonstrated enhanced performance in complex tasks like noise reduction, image reconstruction, and compression artifacts correction. Additionally, this review highlights the challenges in IQA, including the subjectivity of human visual perception and the limitations of various algorithms in handling different types of distortions. It concludes by suggesting future research directions that integrate hybrid models combining classical techniques with deep learning to achieve more robust and efficient image quality evaluation.