Convolutional Deblurring for Natural Imaging

Abstract

In this paper, we propose a novel design of image deblurring in the form of one-shot convolution filtering that can directly convolve with naturally blurred images for restoration. The problem of optical blurring is a common disadvantage to many imaging applications that suffer from optical imperfections. Despite numerous deconvolution methods that blindly estimate blurring in either inclusive or exclusive forms, they are practically challenging due to high computational cost and low image reconstruction quality. Both conditions of high accuracy and high speed are prerequisites for high-throughput imaging platforms in digital archiving. In such platforms, deblurring is required after image acquisition before being stored, previewed, or processed for high-level interpretation. Therefore, on-the-fly correction of such images is important to avoid possible time delays, mitigate computational expenses, and increase image perception quality. We bridge this gap by synthesizing a deconvolution kernel as a linear combination of Finite Impulse Response (FIR) evenderivative filters that can be directly convolved with blurry input images to boost the frequency fall-off of the Point Spread Function (PSF) associated with the optical blur. We employ a Gaussian low-pass filter to decouple the image denoising problem for image edge deblurring. Furthermore, we propose a blind approach to estimate the PSF statistics for two Gaussian and Laplacian models that are common in many imaging pipelines. Thorough experiments are designed to test and validate the efficiency of the proposed method using 2054 naturally blurred images across six imaging applications and seven state-of-the-art deconvolution methods.

Restoration of Blurred and Noisy Images Using Inverse Filtering and Adaptive Threshold Method

“…The proposed scheme in [15] addresses the problem of blind image deconvolution using multiple blurry images algorithm. In [16], the proposed image deblurring method increases the frequency fall-off of the PSF by convolving the blurry input images with a deconvolution kernel composed of a linear combination of finite impulse response (FIR) filters. Estimation of blur kernels is not required in the introduced scheme in [17] using two images acquired in low light.…”

“…The image degradations may include blurring due to camera motion, for example, [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] or noise which is effectively equivalent to errors in the image pixel values and is due to many causes such as electronic image transmission [19][20][21][22][23][24][25][26][27]. Law enforcement, for example, is an application of image restoration in which the image is blurred due to motion.…”

Mobile Computational Photography: A Tour

“…These methods generally combine natural image priors (i.e., what characteristics does a natural sharp image have), and assumptions on the blur kernel (e.g., maximum size) to cast the blind deconvolution problem as one of variational optimization El-Henawy et al ( 2014), Fergus et al (2006), Levin et al (2009). In the specific case of deblurring slightly blurry images, we can proceed in a more direct way by filtering the image with an estimate of the blur and thus avoid using costly optimization procedures Delbracio 2020), Hosseini & Plataniotis (2019). Figure 14 shows an example of Polyblur Delbracio et al (2020) that efficiently removes blur by estimating the blur and combining multiple applications of the estimated blur to approximate its inverse.…”

SlimRGBD: A Geographic Information Photography Noise Reduction System for Aerial Remote Sensing

“…Each fully connected layer is followed by a ReLU layer as an activation function. The global average merge layer is used to ensure that our model can handle images of any resolution [34], [35]. Finally, the global information is summarized as a fixed dimension vector and used to normalize the local features produced by the local path.…”