Constrained Predictive Filters for Single Image Bokeh Rendering

Abstract

Bokeh rendering is a technique used to take pictures with out-of-focus areas to highlight regions of interest. Due to limitations in hardware and shooting condition, rendering a bokeh image from a full-focus image has attracted a lot of interest. In this paper, we model bokeh rendering as the combination of salient region retention and bokeh blurring, and propose a neural network to generate a realistic bokeh image from a single full-focus image through end-to-end training. Specifically, we propose a gate fusion block to estimate the salient area, and introduce a constrained predictive filter for salient region retention and bokeh blurring within a unified architecture. Further, we utilize a pixel coordinate-based map to enhance the training. Experimental results illustrate the effectiveness of our model. The comparison with state-of-the-art methods (PyNET [18], DMSHN [12], BGGAN [40], etc.) shows that our model produces better bokeh effects and retains salient objects.

Quad Bayer Joint Demosaicing and Denoising Based on Dual Encoder Network with Joint Residual Learning

“…Recently, learning-based methods exhibited great potential in image reconstruction tasks (Zheng et al 2022;Chen et al 2023;Zhao et al 2021). (Ehret et al 2019) obtained the processed demosaicing and denoising images by using a convolutional neural network with temporal and spatial redundancy information between consecutively captured images, thereby achieving the joint demosaicing and denoising task in an unsupervised way.…”

Improving Dynamic HDR Imaging with Fusion Transformer

“…However, as the L1 loss is a point-wise loss, it does not capture edge information important particularly to minimize ghosting in the HDR reconstruction. Following the novel training strategy of CNN (Zheng et al 2020(Zheng et al , 2021(Zheng et al , 2022a, we adopt the Advanced Sobel Loss (ASL) and combine it with the L1 loss to formulate the loss function for to enhance the edge information, which can be expressed as:…”