Motion Blur Estimation Research Articles

XAI-FusionNet: Diabetic foot ulcer detection based on multi-scale feature fusion with explainable artificial intelligence

Diabetic foot ulcer (DFU) poses a significant threat to individuals affected by diabetes, often leading to limb amputation. Early detection of DFU can greatly improve the chances of survival for diabetic patients. This work introduces FusionNet, a novel multi-scale feature fusion network designed to accurately differentiate DFU skin from healthy skin using multiple pre-trained convolutional neural network (CNN) algorithms. A dataset comprising 6963 skin images (3574 healthy and 3389 ulcer) from various patients was divided into training (6080 images), validation (672 images), and testing (211 images) sets. Initially, three image preprocessing techniques - Gaussian filter, median filter, and motion blur estimation - were applied to eliminate irrelevant, noisy, and blurry data. Subsequently, three pre-trained CNN algorithms -DenseNet201, VGG19, and NASNetMobile - were utilized to extract high-frequency features from the input images. These features were then inputted into a meta-tuner module to predict DFU by selecting the most discriminative features. Statistical tests, including Friedman and analysis of variance (ANOVA), were employed to identify significant differences between FusionNet and other sub-networks. Finally, three eXplainable Artificial Intelligence (XAI) algorithms - SHAP (SHapley Additive exPlanations), LIME (Local Interpretable Model-agnostic Explanations), and Grad-CAM (Gradient-weighted Class Activation Mapping) - were integrated into FusionNet to enhance transparency and explainability. The FusionNet classifier achieved exceptional classification results with 99.05 % accuracy, 98.18 % recall, 100.00 % precision, 99.09 % AUC, and 99.08 % F1 score. We believe that our proposed FusionNet will be a valuable tool in the medical field to distinguish DFU from healthy skin.

Self-calibration method for rotating Risley prism main sections based on chromatic dispersion characteristics of the Fourier spectrum.

To address the main section misalignment of rotating Risley prisms in a camera imaging system, a zero calibration method for a Risley prism main section is proposed according to the dispersion characteristics of the prisms-camera imaging. The principal component analysis method and the Radon transform method are combined to establish an effective mapping relationship between the prism main section positions and the prismatic dispersion characteristics using the camera dispersion image as references. The estimation of the dispersion direction was realized by referring to the motion blur estimation method and verified by experiment. As a result, the target-free self-calibration of the prism main sections in the camera coordinate system can be carried out under a prisms-camera intact assembly mode.

RFID tag group recognition based on motion blur estimation and YOLOv2 improved by Gaussian

RFID tag group recognition based on motion blur estimation and YOLOv2 improved by Gaussian

Kernel learning for blind image recovery from motion blur

Restoring image from motion deblur faces great challenges in the estimation of the motion blur kernel that is the key to recover the latent sharp image. In this paper, we present a method to iteratively estimate the structural image and account for the textural component. A scale-aware smoothing operation is developed to remove fine-scale edges with resampling. Our method leverages L0-norm regularization to enforce the sparsity of the motion blur kernel in both intensity and derivative domains. Experiments are conducted to evaluate the performance of our proposed method using two widely accepted public datasets. We found that our proposed method is insensitive to most hyper-parameters. Both qualitative evaluation and quantitative evaluation confirms that our method effectively restores the sharp image without introducing artifacts. The minimum improvements in terms of average PSNR for both datasets are more than 3.13% for all cases and the improvements in terms of average error rate are 15%. By visually comparing the estimated motion blur kernels, it is clear that the estimated kernel by our method is the closest to the actual kernel used to generate the synthesized blurry images.

Learning to Deblur Face Images via Sketch Synthesis

The success of existing face deblurring methods based on deep neural networks is mainly due to the large model capacity. Few algorithms have been specially designed according to the domain knowledge of face images and the physical properties of the deblurring process. In this paper, we propose an effective face deblurring algorithm based on deep convolutional neural networks (CNNs). Motivated by the conventional deblurring process which usually involves the motion blur estimation and the latent clear image restoration, the proposed algorithm first estimates motion blur by a deep CNN and then restores latent clear images with the estimated motion blur. However, estimating motion blur from blurry face images is difficult as the textures of the blurry face images are scarce. As most face images share some common global structures which can be modeled well by sketch information, we propose to learn face sketches by a deep CNN so that the sketches can help the motion blur estimation. With the estimated motion blur, we then develop an effective latent image restoration algorithm based on a deep CNN. Although involving the several components, the proposed algorithm is trained in an end-to-end fashion. We analyze the effectiveness of each component on face image deblurring and show that the proposed algorithm is able to deblur face images with favorable performance against state-of-the-art methods.

Blur Length Estimation Based on Image Energy Spectrum Statistics

In order to resolve the non-smooth blur problem for coded exposure imaging, this paper proposes an automatically motion blur estimation method based on image energy spectrum statistics, and expands successfully this method to estimate motion blur length for traditional motion blurred images. This paper finds that the residual sums of squares of polynomial fitness result of the deblurred image energy spectrum statistics can reach the minimum, only using correct blur length in the deconvolution process. Given an initial value, using simulated annealing algorithm to implement the iteration, the correct blur length can be obtained automatically in a very short time. Both the real and simulated experiment results demonstrate the effectiveness and robustness of the proposed method.

Blind Image Deconvolution using Frequency Spectrum in Motion Blur Estimation

Image deconvolution refers to the act of minimizing the blur quantity from a blurred image and discovers the sharp and clear original image [7].The blind image deconvolution (BID), refers to the task of separating two convolved signals.We have taken motion blur estimation method which will extract the features from the frequency domain of the image. This method has low complexity as compare to other estimation methods like radon, steerable filter methods.

Model-based motion blur estimation for the improvement of motion tracking

Video tracking is an important task in many automated or semi-automated applications, like cinematic post production, surveillance or traffic monitoring. Most established video tracking methods fail or lead to an inaccurate estimate when motion blur occurs in the video, as they assume, that the object appears constantly sharp in the video. In this paper, we present a novel motion tracking method with explicit modeling of motion blur, estimating the continuous motion of a rigid 3-D object with known geometry in a monocular video as well as the sharp object texture. Instead of treating motion blur as a potential source of errors, we take advantage of it and consider motion blur as an additional information source, providing information about the motion of the tracked object during the exposure. In an analysis-by-synthesis approach we explicitly model the effects of motion blur reconstructing the captured frames, in order to accomplish a more accurate estimation. We design our algorithm to be capable to run in parallel on the GPU using the common rendering pipeline and considering each frame individually to handle also long videos. We tested our approach on both synthetic and real videos. In both cases, we achieve significant improvements of accuracy and reductions of frame reconstruction error compared to the estimated motion of a rigid body tracker, without motion blur handling.

Motion blur estimation based on multitarget matching model

We propose a new method to estimate motion blur parameters based on the autocorrelation function of a blurred image. This blurred image is considered as a superposition of M shifted images identical to the original nonblurred image. In this case, convolution of the blurred image with itself can be considered as M 2 pairwise convolutions, which contribute to the resultant autocorrelation function producing a distinguishable line corresponding to the estimated motion blur angle. The proposed method demonstrates the comparable accuracy of the motion blur angle estimation in comparison with state-of-the-art methods. Our method possesses lower computational complexity than popular accurate methods based on Radon transform. The proposed model also allows to accurately estimate motion blur length. Our results of length estimation, in general, outperform the accuracy of the methods based on Radon transform.

Blind motion image deblurring using nonconvex higher-order total variation model

We propose a nonconvex higher-order total variation (TV) method for blind motion image deblurring. First, we introduce a nonconvex higher-order TV differential operator to define a new model of the blind motion image deblurring, which can effectively eliminate the staircase effect of the deblurred image; meanwhile, we employ an image sparse prior to improve the edge recovery quality. Second, to improve the accuracy of the estimated motion blur kernel, we use L1 norm and H1 norm as the blur kernel regularization term, considering the sparsity and smoothing of the motion blur kernel. Third, because it is difficult to solve the numerically computational complexity problem of the proposed model owing to the intrinsic nonconvexity, we propose a binary iterative strategy, which incorporates a reweighted minimization approximating scheme in the outer iteration, and a split Bregman algorithm in the inner iteration. And we also discuss the convergence of the proposed binary iterative strategy. Last, we conduct extensive experiments on both synthetic and real-world degraded images. The results demonstrate that the proposed method outperforms the previous representative methods in both quality of visual perception and quantitative measurement.

An Automated Algorithm for Measurement of Surgical Tip Excursion in Ultrasonic Vibration Using the Spatial 2-Dimensional Fourier Transform in an Optical Image

The International Electrotechnical Commission (IEC) has defined a standard IEC 61847 (First Edition, 1998) for characterization of ultrasonic surgical systems. This standard prescribes several methods for measurement of primary tip vibration excursion. The first method described in the standard uses an optical microscope and relies on the motion blur of a vibrating object as it is imaged at low frame rates (e.g. 30 Hz) of conventional video equipment. This is a widely used method, that predates the standard, in ultrasonic surgical instrument design, and it is one of the key parameters that surgeons who use these devices are aware of. It is relatively easily measured using a microscope system. Although this method is widespread, the accuracy of this method is highly dependent on multiple factors such as operator training, microscope lighting and modulation of surgical tip motion. It is also a manual and time consuming measurement such that a continuous measurement that describes dynamics at the scale of micro-seconds becomes impossible. Here we describe an algorithm to automate this measurement so that it can be done at high speed without operator training, reducing human error and operator variation. The algorithm derives from techniques used in motion blur estimation and reduction in the image processing literature. A 2 dimensional spatial Fourier transform is computed from the microscope image of an ultrasonically vibrating tip. A peak detection algorithm is used along with pre-processing to reduce noise. Separation of peaks in the Fourier domain is used to estimate tip excursion. We present data that shows an error of about 1% between manual and automated methods, when measurements are in the range of 300 microns and about 20% when the measurements are in the range of 30 microns.

Noncausal fractional directional differentiator and blind deconvoluation: motion blur estimation

The problem of blind estimation of motion blur parameters from a single image is addressed. The blur direction and extent of motion-blurred image, which are introduced by relative motion between a camera and its object scene, are needed in the methods of image restoration, such as blind deconvolution. As an extension to the fractional-order derivative, a noncausal fractional-order directional derivative operator is devised, which is robust to noise. Based on this new operator, a novel method identifying blur parameters is developed in this work. The performance comparison between the proposed method and the state-of-the-art method is also presented, demonstrating that the former provides better immunity to noise and capacity to identify motion blur extent, especially for large blur length.

PSF Estimation and Image Restoration for Motion Blurred Images

Point spread function (PSF) estimation and image restoration algorithm are the hotspots In the research of motion blurred image restoration. In order to improve the efficacy of image restoration, an improved algorithm named quadric transforms (QT) method is proposed in this paper by analyzing the restoration process of motion blurred images. Firstly, Fourier transform and homomorphism transform are applied to the original motion blurred image, and then the Fourier transform and homomorphism transform are used again to the obtained spectrum image. Secondly, the motion blur direction is estimated by Radon transform. Thirdly, the motion blur length is found by differential autocorrelation operations. Finally, utilizing the estimated blur direction and blur length, the motion blurred image is restored by Wiener filtering. Experimental results show that the proposed QT method can get more accurate estimated motion blur angles than the primary transform (PT, that is, Fourier transform and homomorphism transform are used one time) method and can get better restored images under the meaning of peak signal to noise ratio (PSNR).

Handling Motion-Blur in 3D Tracking and Rendering for Augmented Reality

The contribution of this paper is two-fold. First, we show how to extend the ESM algorithm to handle motion blur in 3D object tracking. ESM is a powerful algorithm for template matching-based tracking, but it can fail under motion blur. We introduce an image formation model that explicitly consider the possibility of blur, and shows its results in a generalization of the original ESM algorithm. This allows to converge faster, more accurately and more robustly even under large amount of blur. Our second contribution is an efficient method for rendering the virtual objects under the estimated motion blur. It renders two images of the object under 3D perspective, and warps them to create many intermediate images. By fusing these images we obtain a final image for the virtual objects blurred consistently with the captured image. Because warping is much faster than 3D rendering, we can create realistically blurred images at a very low computational cost.

Exposing Digital Image Forgeries by Detecting Discrepancies in Motion Blur

The widespread availability of photo manipulation software has made it unprecedentedly easy to manipulate images for malicious purposes. Image splicing is one such form of tampering. In recent years, researchers have proposed various methods for detecting such splicing. In this paper, we present a novel method of detecting splicing in images, using discrepancies in motion blur. We use motion blur estimation through image gradients in order to detect inconsistencies between the spliced region and the rest of the image. We also develop a new measure to assist in inconsistent region segmentation in images that contain small amounts of motion blur. Experimental results show that our technique provides good segmentation of regions with inconsistent motion blur. We also provide quantitative comparisons with other existing blur-based techniques over a database of images. It is seen that our technique gives significantly better detection results.