Abstract:The limited depth of field (DOF) inherent in cameras often results in defocused and blurry backgrounds when capturing images in large aperture mode. This not only leads to the loss of crucial background information but also impedes the efficient reconstruction of the background regions. Usually, super-resolution (SR) techniques struggle to produce high-quality results for images captured with large apertures. To enhance the reconstruction quality of defocused regions in large aperture images, a foreground-background separation and deblurring super-resolution (FBSDSR) method was proposed in this paper. Based on the idea of foreground-background separation processing, we first divide the large aperture image into a sharp foreground region (If) and a blurry background region (Ib) based on depth information. The background region (Ib) is then deblurred using an end-to-end iterative filter adaptive network (IFAN). This deblurring process refocuses the background, ultimately restoring an image with sharp details throughout. Finally, the enhanced super-resolution generative adversarial networks (Real-ESRGAN) which specializes in images SR of realistic scenes was used to process the sharp all-in-focus image. This method results in high-quality reconstructions of both the foreground and background of large aperture images. The experimental results demonstrated that the proposed method achieved effective reconstruction of the entire large aperture images clearly, overcoming the limitations of existing methods that struggle to reconstruct defocused regions. This significantly enhances the quality and resolution of large aperture images. Specifically, when FBSDSR is integrated with Real-ESRGAN, the PSNR, LPIPS, NIQE, and hyperIQA metrics were improved by approximately 2.2 %, 45.1 %, 34.7 %, and 10.9 % respectively.

Multi-frame image super-resolution reconstruction based on spatial information weighted fields of experts

Multi-frame Image Super Resolution with Natural Image Prior

A Multi-Frame Image Super-Resolution Method.

A Super-Resolution Method with EWA

Investigation of Multi-Frame Image Super-Resolution Based on Adaptive Self-Learning Methods

A Single-Image Super-Resolution Method Based on Progressive-Iterative Approximation

A novel approach to multi-frame image super resolution using an innovative filter and pixel selection mechanism

Multi-frame Super Resolution Using Refined Exploration of Extensive Self-examples.

A weighted least squares algorithm for time-of-flight depth image denoising

Image super-resolution using nonlocally centralized sparse representation and fields of experts priors

A noise-suppressing and edge-preserving multiframe super-resolution image reconstruction method

Multi-sensor Image Super-Resolution with Fuzzy Cluster by Using Multi-Scale and Multi-View Sparse Coding for Infrared Image

Single image super resolution via wavelet transform fusion and SRFeat network

Multi-frame Image Super-Resolution Based on Knife-Edges

Multi-Stage Degradation Homogenization for Super-Resolution of Face Images With Extreme Degradations

Single-Frame Image Super-Resolution Inspired by Perceptual Criteria

Single-image super-resolution using directional total variation regularization and alternating direction method of multiplier solver

Multi-Stream Fusion Network for Multi-Distortion Image Super-Resolution

Foreground-background Separation and Deblurring Super-Resolution Method

Robust Remote Sensing Super-Resolution With Frequency Domain Decoupling for Multiscenarios

Multi-scale Single Image Super-Resolution with Remote-Sensing Application Using Transferred Wide Residual Network