Defocus Research Articles

Peripheral Choroidal Response to Localized Defocus Blur: Influence of Native Peripheral Aberrations.

This study aims to examine the short-term peripheral choroidal thickness (PChT) response to signed defocus blur, both with and without native peripheral aberrations. This examination will provide insights into the role of peripheral aberration in detecting signs of defocus. The peripheral retina (temporal 15°) of the right eye was exposed to a localized video stimulus in 11 young adults. An adaptive optics system induced 2D myopic or hyperopic defocus onto the stimulus, with or without correcting native peripheral ocular aberrations (adaptive optics [AO] or NoAO defocus conditions). Choroidal scans were captured using Heidelberg Spectralis OCT at baseline, exposure (10, 20, and 30minutes), and recovery phases (4, 8, and 15minutes). Neural network-based automated MATLAB segmentation program measured PChT changes from OCT scans, and statistical analysis evaluated the effects of different optical conditions over time. During the exposure phase, NoAO myopic and hyperopic defocus conditions exhibited distinct bidirectional PChT alterations, showing average thickening (10.0 ± 5.3µm) and thinning (-9.1 ± 5.5µm), respectively. In contrast, induced AO defocus conditions did not demonstrate a significant change from baseline. PChT recovery to baseline occurred for all conditions. The unexposed fovea did not show any significant ChT change, indicating a localized ChT response to retinal blur. We discovered that the PChT response serves as a marker for detecting peripheral retinal myopic and hyperopic defocus blur, especially in the presence of peripheral aberrations. These findings highlight the significant role of peripheral oriented blur in cueing peripheral defocus sign detection.

Dynamic control of defocus, astigmatism, and tilt aberrations with a large area foveal liquid crystal lens

We have recently reported the dynamic adjustment of the focal length in an electrically tunable liquid crystal “foveal” lens, the center of which can be shifted over a large working area. In the present work, we show that this design allows also the independent generation of astigmatism with arbitrary axis and tilt of the light wavefront by simply changing the phase and the voltage differences between 4 control electrodes. Furthermore, we also demonstrate the capability of generating highly localized negative (defocusing) lenses with the same device by using a dual frequency liquid crystal.

Prior and Prediction Inverse Kernel Transformer for Single Image Defocus Deblurring

Defocus blur, due to spatially-varying sizes and shapes, is hard to remove. Existing methods either are unable to effectively handle irregular defocus blur or fail to generalize well on other datasets. In this work, we propose a divide-and-conquer approach to tackling this issue, which gives rise to a novel end-to-end deep learning method, called prior-and-prediction inverse kernel transformer (P2IKT), for single image defocus deblurring. Since most defocus blur can be approximated as Gaussian blur or its variants, we construct an inverse Gaussian kernel module in our method to enhance its generalization ability. At the same time, an inverse kernel prediction module is introduced in order to flexibly address the irregular blur that cannot be approximated by Gaussian blur. We further design a scale recurrent transformer, which estimates mixing coefficients for adaptively combining the results from the two modules and runs the scale recurrent ``coarse-to-fine" procedure for progressive defocus deblurring. Extensive experimental results demonstrate that our P2IKT outperforms previous methods in terms of PSNR on multiple defocus deblurring datasets.

High quality holographic 3D display with enhanced focus cues based on multiple directional light reconstruction.

Holographic display faces the trade-off between image quality and focus cues, resulting from the specific choice of phase distribution. In this Letter, we propose a speckle-free holographic display with enhanced focus cues by multiple directional light reconstruction. The uniform phase hologram is first generated by the gradient descent optimization algorithm. The blazed grating phase is used to steer the object light to a specific direction. Multiple sub-holograms with different blazed gratings are refreshed fast to reconstruct the images from different directions. Thus, the defocus blur is improved due to the separation of multiple spots on the defocus plane. The multi-plane reconstruction is also realized by pre-processing the depth images to eliminate image separation. The proposed method provides apparent focus cues while maintaining high image qualities, which is expected to realize comfortable holographic near-eye display in the near future.

Improved modeling of human vision by incorporating robustness to blur in convolutional neural networks

Whenever a visual scene is cast onto the retina, much of it will appear degraded due to poor resolution in the periphery; moreover, optical defocus can cause blur in central vision. However, the pervasiveness of blurry or degraded input is typically overlooked in the training of convolutional neural networks (CNNs). We hypothesized that the absence of blurry training inputs may cause CNNs to rely excessively on high spatial frequency information for object recognition, thereby causing systematic deviations from biological vision. We evaluated this hypothesis by comparing standard CNNs with CNNs trained on a combination of clear and blurry images. We show that blur-trained CNNs outperform standard CNNs at predicting neural responses to objects across a variety of viewing conditions. Moreover, blur-trained CNNs acquire increased sensitivity to shape information and greater robustness to multiple forms of visual noise, leading to improved correspondence with human perception. Our results provide multi-faceted neurocomputational evidence that blurry visual experiences may be critical for conferring robustness to biological visual systems.

Deblurring using Reverse Filtering Techniques: An Experimental Analysis

Images are an integral part of life. Blurring an image is a process that makes the image less sharp and distorts the detail of an image. The images got blurred due to atmospheric noise, improper camera setting, and relative motion between the camera and the scene. Deblurring is the process of recovering the sharp image from the blurred image. The types of blurs include Average blur, Motion blur, Defocus blur, Gaussian blur, etc. Mainly deblurring techniques are categorised into blind image deconvolution and nonblind image deconvolution. There exist many deblurring techniques in the literature. Deblurring using reverse filtering is to suppress the effect of the filter using the filter itself as many times as needed but without knowing its internal structure. The filter is regarded as a black box. This paper presents the experimental analysis of existing reverse filtering techniques such as zero order reverse filtering (T-method), Rendition method (R-method), P-method, and TDA (Total Derivative Approximation). The methods considered in the paper are spatial domain methods.

Defocus blur detection via adaptive cross-level feature fusion and refinement

Defocus blur detection via adaptive cross-level feature fusion and refinement

Head-mounted adaptive optics visual simulator.

Adaptive optics visual simulation is a powerful tool for vision testing and evaluation. However, the existing instruments either have fixed tabletop configurations or, being wearable, only offer the correction of defocus. This paper proposes a novel head-mounted adaptive optics visual simulator that can measure and modify complex ocular aberrations in real-time. The prototype is composed of two optical modules, one for the objective assessment of aberrations and the second for wavefront modulation, all of which are integrated into a wearable headset. The device incorporates a microdisplay for stimulus generation, a liquid crystal on silicon (LCoS) spatial light modulator for wavefront manipulation, and a Hartmann-Shack wavefront sensor. Miniature optical components and optical path folding structures, together with in-house 3D printed mounts and housing, were adapted to realize the compact size. The system was calibrated by characterizing and compensating the internal aberrations of the visual relay. The performance of the prototype was analyzed by evaluating the measurement and compensation of low-order and higher-order aberrations induced through trial lenses and phase masks in an artificial eye. The defocus curves for a simulated bifocal diffractive lens were evaluated in real eyes. The results show high accuracy while measuring and compensating for the induced defocus, astigmatism, and higher-order aberrations, whereas the MTF analysis shows post-correction resolution of up to 37.5 cycles/degree (VA 1.25). Moreover, the subjective test results show the defocus curves closely matched to a commercial desktop visual simulator.

Numerical study of satellite image quality in the presence of Seidel aberrations

AbstractWe present estimated images of a hypothetical satellite system in the presence of Seidel aberrations. We exemplified the effects of defocus, spherical, coma, astigmatism, field curvature, and distortion aberrations. Our analysis methods allow the investigation of aberration effects individually and in any combined form. Thereby, we introduce a potentially rich analysis dimension for the imaging chain, image processing, and image quality analyses. For a number of degrees of aberrations, we provide image estimates, show and quantify how individual (field‐dependent) aberrations would introduce degradations on the satellite images. We got correlation coefficient values for defocus, spherical, coma, astigmatism, field curvature, and distortion aberrations as 0.69, 0.71, 0.66, 0.79, 0.76, and 0.60, respectively. Thereby, we identfied the significance order of Seidel aberrations for this system by using image simulations and image similarity measures.

Efficient Distortion-Free Neural Projector Deblurring in Dynamic Projection Mapping.

Dynamic Projection Mapping (DPM) necessitates geometric compensation of the projection image based on the position and orientation of moving objects. Additionally, the projector's shallow depth of field results in pronounced defocus blur even with minimal object movement. Achieving delay-free DPM with high image quality requires real-time implementation of geometric compensation and projector deblurring. To meet this demand, we propose a framework comprising two neural components: one for geometric compensation and another for projector deblurring. The former component warps the image by detecting the optical flow of each pixel in both the projection and captured images. The latter component performs real-time sharpening as needed. Ideally, our network's parameters should be trained on data acquired in an actual environment. However, training the network from scratch while executing DPM, which demands real-time image generation, is impractical. Therefore, the network must undergo pre-training. Unfortunately, there are no publicly available large real datasets for DPM due to the diverse image quality degradation patterns. To address this challenge, we propose a realistic synthetic data generation method that numerically models geometric distortion and defocus blur in real-world DPM. Through exhaustive experiments, we have confirmed that the model trained on the proposed dataset achieves projector deblurring in the presence of geometric distortions with a quality comparable to state-of-the-art methods.

Voice Vista

Blind image restoration involves leveraging prior information within an image to restore the sharpness of edges. De-blurring, on the other hand, aims to eliminate blurring artifacts caused by factors like defocus aberration or motion blur, which manifests as apparent streaking in still images of rapidly moving objects. Gaussian blur results from applying a Gaussian function to blur an image. Recent advancements in single-image methods owe their success in part to the utilization of various sparse priors for either latent images or motion blur kernels. KSR introduces an efficient kernel matrix approximation to accelerate blurring processes and achieve notable de-blur performance on digital datasets. License plate recognition serves as a critical tool for identifying over-speed vehicles or those involved in hit-and-run incidents. However, surveillance camera snapshots of speeding vehicles often suffer from motion blur, rendering them unrecognizable to the human eye and presenting a challenge to existing blind deblurring techniques. To address this, we propose a novel sparse representation-based scheme for identifying motion blur kernels. By analyzing sparse representation coefficients of the recovered image, we determine the kernel angle, and estimate the length of the motion kernel using Radon transform in Fourier domain. Our approach effectively handles large motion blur, even when license plates are unidentifiable by humans. Experimental evaluations on real-world images demonstrate the superiority of our method over several state-of-the-art blind image de-blurring algorithms.

Functional changes in the myopic retina compromise emmetropization

During emmetropization, the retina analyses the focus of the image and generates appropriate biochemical signals to control choroidal thickness and scleral growth. Using commercially available optical low coherence interferometry (the Lenstar LS 900), changes in eye length in the micrometre range can be picked up already after 30 min when young human subjects watch movies that were digitally filtered. We found that the emmetropic retina can distinguish real optical positive defocus from simulated/calculated defocus. It triggered choroidal thickening in the first case, and thinning in the second. Strikingly, the myopic retina was not responsive, and the choroid became thinner in both cases. Longitudinal chromatic aberration (LCA) would be a perfect signal to tell the retina where the focal plane is. We simulated chromatic defocus in our movies. Either the red or the blue plane of the RGB format was spatially filtered to simulate chromatic defocus. When blue was ‘in focus’, eyes became longer, but when red was ‘in focus’ eyes became shorter. Again, myopic eyes did not respond, suggesting that the myopic retina no longer responded to LCA. Taken together, we found that the myopic retina had undergone functional changes such that no growth‐inhibiting signal was generated. As a consequence, eye growth would become open loop, explaining why myopia does not inhibit itself.

Incrementally Adapting Pretrained Model Using Network Prior for Multi-Focus Image Fusion.

Multi-focus image fusion can fuse the clear parts of two or more source images captured at the same scene with different focal lengths into an all-in-focus image. On the one hand, previous supervised learning-based multi-focus image fusion methods relying on synthetic datasets have a clear distribution shift with real scenarios. On the other hand, unsupervised learning-based multi-focus image fusion methods can well adapt to the observed images but lack the general knowledge of defocus blur that can be learned from paired data. To avoid the problems of existing methods, this paper presents a novel multi-focus image fusion model by considering both the general knowledge brought by the supervised pretrained backbone and the extrinsic priors optimized on specific testing sample to improve the performance of image fusion. To be specific, the Incremental Network Prior Adaptation (INPA) framework is proposed to incrementally integrate features extracted from the pretrained strong baselines into a tiny prior network (6.9% parameters of the backbone network) to boost the performance for test samples. We evaluate our method on both synthetic and real-world public datasets (Lytro, MFI-WHU, and Real-MFF) and show that our method outperforms existing supervised learning-based methods and unsupervised learning based methods.

Multi-Interactive Enhanced for Defocus Blur Estimation

Multi-Interactive Enhanced for Defocus Blur Estimation

Deep Depth from Focal Stack with Defocus Model for Camera-Setting Invariance

We propose deep depth from focal stack (DDFS), which takes a focal stack as input of a neural network for estimating scene depth. Defocus blur is a useful cue for depth estimation. However, the size of the blur depends on not only scene depth but also camera settings such as focus distance, focal length, and f-number. Current learning-based methods without any defocus models cannot estimate a correct depth map if camera settings are different at training and test times. Our method takes a plane sweep volume as input for the constraint between scene depth, defocus images, and camera settings, and this intermediate representation enables depth estimation with different camera settings at training and test times. This camera-setting invariance can enhance the applicability of DDFS. The experimental results also indicate that our method is robust against a synthetic-to-real domain gap.

Telephoto achromatic camera based on optical–digital co-design

Due to the difficulty of correcting chromatic aberration (CA) in telephoto cameras, recent studies have combined image algorithms with simple optical structures, such as single-spherical lenses, for high-quality photography, moving away from complex optics. However, this approach often struggles to comprehensively address compounded issues arising from optical aberrations of simple optical systems, including defocus blur and multi-channel misalignment. To tackle this challenge, this manuscript presents an approach for developing a telephoto imaging system by leveraging the distinct characteristics of axial and lateral chromatic aberrations (ACA, LCA) over the visible spectrum. The optical design is limited to a specific wavelength range to preserve high-frequency information of the green channel. A cross-channel fitting method is presented to suppress the LCA. Subsequently, the powerful capabilities of deep learning are utilized to correct ACA, defocus blur, and other residual optical aberrations. Simulation experiments demonstrate the effectiveness of the proposed approach in mitigating the CA inherent in telephoto systems, thereby delivering high-quality imaging results over the whole visible waveband.

An Implicit Neural Representation for the Image Stack: Depth, All in Focus, and High Dynamic Range

In everyday photography, physical limitations of camera sensors and lenses frequently lead to a variety of degradations in captured images such as saturation or defocus blur. A common approach to overcome these limitations is to resort to image stack fusion, which involves capturing multiple images with different focal distances or exposures. For instance, to obtain an all-in-focus image, a set of multi-focus images is captured. Similarly, capturing multiple exposures allows for the reconstruction of high dynamic range. In this paper, we present a novel approach that combines neural fields with an expressive camera model to achieve a unified reconstruction of an all-in-focus high-dynamic-range image from an image stack. Our approach is composed of a set of specialized implicit neural representations tailored to address specific sub-problems along our pipeline: We use neural implicits to predict flow to overcome misalignments arising from lens breathing, depth, and all-in-focus images to account for depth of field, as well as tonemapping to deal with sensor responses and saturation - all trained using a physically inspired supervision structure with a differentiable thin lens model at its core. An important benefit of our approach is its ability to handle these tasks simultaneously or independently, providing flexible post-editing capabilities such as refocusing and exposure adjustment. By sampling the three primary factors in photography within our framework (focal distance, aperture, and exposure time), we conduct a thorough exploration to gain valuable insights into their significance and impact on overall reconstruction quality. Through extensive validation, we demonstrate that our method outperforms existing approaches in both depth-from-defocus and all-in-focus image reconstruction tasks. Moreover, our approach exhibits promising results in each of these three dimensions, showcasing its potential to enhance captured image quality and provide greater control in post-processing.

Investigation of blur kernel of terahertz images

The paper discusses issues of digital processing of terahertz images. It is shown that despite the improvement of the hardware part of imaging setups, the acquired images still often have a low resolution and suffer from noise and blurring effects. Thus, to improve their visual quality, it is advisable to use special digital processing methods. While some progress has already been made in terms of denoising of terahertz images, the research of their deblurring is only at the very early stage. Therefore, this paper attempts to analyze the properties of blur functions for real terahertz images to further use them while designing a corresponding deblurring technique. For this purpose, the phase-only image method has been used. A study of blur properties for the three most common blur types (defocus, motion and Gaussian blur) has shown that for test images they can be distinguished and their main parameters can be assessed. However, the application of this method to real terahertz images has shown that the blur characteristics in them are very different from the ones obtained for modelled examples. The real blur demonstrates a quite complex behaviour and estimating its kernel requires additional research.

Effect of wearing peripheral focus-out glasses on emmetropization in Chinese children aged 6–8 years: study protocol for a 2-year randomized controlled intervention trial

BackgroundMyopia is one of the most common eye diseases causing visual impairment and blindness, and the high prevalence in adolescents remains a major public health concern. Based on clinical studies using optical defocus to regulate ocular growth and refractive changes through visual feedback, we hypothesize that early wearing of peripheral myopic defocusing spectacles in children with high risk of myopia may slow the process of emmetropization and even prevent the onset of myopia by inducing more peripheral myopic defocus. The aim of this study is to investigate whether the wearing of peripheral focus-out glasses can be effective in delaying emmetropization in non-myopic children aged 6–8 years.MethodsThe study is a 2-year randomized controlled trial. A total of 160 subjects will be randomized into the experimental group or the control group. The experimental group will be fitted with direct emmetropia with focus-out glasses (DEFOG) to guide the emmetropization process. The control group will not receive any treatment and will serve as a blank control group. The primary aim is to determine whether non-myopic children wearing DEFOG lenses are superior to those who do not receive any intervention on the progression of cycloplegic objective refraction over 2 years.DiscussionThis is the first randomized controlled trial aiming at myopic prevention by non-invasive intervention in non-myopic children. This study aims to initially investigate whether wearing peripheral focus-out glasses can effectively delay the process of emmetropization in children aged 6–8 years with high risk of myopia, which might give potential clues for further exploration on early prevention of myopia.Trial registrationClinicalTrials.gov NCT05689567. Registered on 10 January 2023.

FPGA-Based CNN for Eye Detection in an Iris Recognition at a Distance System

Neural networks are the state-of-the-art solution to image-processing tasks. Some of these neural networks are relatively simple, but the popular convolutional neural networks (CNNs) can consist of hundreds of layers. Unfortunately, the excellent recognition accuracy of CNNs comes at the cost of very high computational complexity, and one of the current challenges is managing the power, delay and physical size limitations of hardware solutions dedicated to accelerating their inference process. In this paper, we describe the embedding of an eye detection system on a Zynq XCZU4EV UltraScale+ multiprocessor system-on-chip (MPSoC). This eye detector is used in the application framework of a remote iris recognition system, which requires high resolution images captured at high speed as input. Given the high rate of eye regions detected per second, it is also important that the detector only provides as output images eyes that are in focus, discarding all those seriously affected by defocus blur. In this proposal, the network will be trained only with correctly focused eye images to assess whether it can differentiate this pattern from that associated with the out-of-focus eye image. Exploiting the neural network’s advantage of being able to work with multi-channel input, the inputs to the CNN will be the grey level image and a high-pass filtered version, typically used to determine whether the iris is in focus or not. The complete system synthetises other cores and implements CNN using the so-called Deep Learning Processor Unit (DPU), the intellectual property (IP) block released by AMD/Xilinx. Compared to previous hardware designs for implementing FPGA-based CNNs, the DPU IP supports extensive deep learning core functions, and developers can leverage DPUs to conveniently accelerate CNN inference. Experimental validation has been successfully addressed in a real-world scenario working with walking subjects, demonstrating that it is possible to detect only eye images that are in focus. This prototype module includes a CMOS digital image sensor that provides 16 Mpixel images, and outputs a stream of detected eyes as 640 × 480 images. The module correctly discards up to 95% of the eyes present in the input images as not being correctly focused.