Advancements in Communication and Systems
Exploring Advanced Deep Learning Techniques for Real-World Image Deblurring
Authors: Rohini Ashok Bhadane and Amol D. Potgantwar
Publishing Date: 13-09-2024
ISBN: 978-81-955020-7-3
Abstract
Real-world image blur is a common problem in various imaging applications, and traditional approaches often fail to produce satisfactory results. The use of deep learning has completely altered the picture processing, and real blur removal is a crucial application of this technology. The advent applications of deep learning have unveiled fresh new opportunities for addressing this challenge. The review shows a comprehensive comparison for various techniques based on deep learning, to remove complex blur. The review compares the techniques based on several parameters, such as performance metrics, computation time, and dataset requirements. Additionally, the review offers an insight into the strengths as well as weaknesses of each technique and identifies future research areas. The study also uses criteria such as the Structural Similarity Index (SSIM), Mean Square Error (MSE), peak signal-to-noise ratio (PSNR), to assess how well these methods work. and visual quality. The results show that deep learning-based approaches offer a promising solution for real-world blur removal, with significant improvements in both quantitative and qualitative metrics.
Keywords
Real-World Blur, Space Variant Blur, Space Invariant Blur, Image Deblurring.
Cite as
Rohini Ashok Bhadane and Amol D. Potgantwar, "Exploring Advanced Deep Learning Techniques for Real-World Image Deblurring", In: Ashish Kumar Tripathi and Vivek Shrivastava (eds), Advancements in Communication and Systems, SCRS, India, 2024, pp. 635-651. https://doi.org/10.56155/978-81-955020-7-3-56