Search

Image upscaling aims to increase the resolution and size of a low resolution (LR) image in order to generate a high resolution (HR) image of high frequency (HF). There are several polynomial methods for obtaining a sharpened, upscaled HR image. The interpolated pixel is measured using a weighted average of the neighboring pixels within the image grid that blur at HF regions in these methods. Edge degradation is also caused by other edge-directed and learning-based upscaling methods, which produce blurring artifacts. A novel approach is proposed to fill these gaps. Using the concept of unsharp masking (USM), the LR image is blurred adaptively based on the region’s local variance. The sharpened high pass filtered (HPF) image is then obtained by subtracting the adaptively blurred image from the LR image. According to USM, the HPF image is combined with the LR image via a gain factor optimized using the cuckoo search (CS) algorithm. To compensate for the loss caused by upscaling, this pre-processing step is performed prior to interpolation. Aside from that, the edge of the B-spline interpolated image is detected and expanded. Edge expansion of the upscaled image is performed to further restore the HF details and reduce zigzag artifacts introduced by upscaling while also preserving the edge boundary. The proposed method outperforms the Lanczos, Bicubic, and Bilinear schemes in terms of peak signal to noise ratio (PSNR) gain of 3.475, 8.3839, and 8.075 dB, respectively. In terms of performance, this method outperforms state-of-the-art techniques both objectively (PSNR and SSIM) and subjectively (visual quality).

Most existing face image Super‐Resolution (SR) methods assume that the Low‐Resolution (LR) images were artificially downsampled from High‐Resolution (HR) images with bicubic interpolation. This operation changes the natural image characteristics and reduces noise. Hence, SR methods trained on such data most often fail to produce good results when applied to real LR images. To solve this problem, a novel framework for the generation of realistic LR/HR training pairs is proposed. The framework estimates realistic blur kernels, noise distributions, and JPEG compression artifacts to generate LR images with similar image characteristics as the ones in the source domain. This allows to train an SR model using high‐quality face images as Ground‐Truth (GT). For better perceptual quality, a Generative Adversarial Network (GAN) based SR model is used, where the commonly used VGG‐loss [1] is exchanged with LPIPS‐loss [2]. Experimental results on both real and artificially corrupted face images show that our method results in more detailed reconstructions with less noise compared to the existing State‐of‐the‐Art (SoTA) methods. In addition, it is shown that the traditional non‐reference Image Quality Assessment (IQA) methods fail to capture this improvement and demonstrate that the more recent NIMA metric [3] correlates better with human perception via Mean Opinion Rank (MOR).

Single image super-resolution (SISR), as a traditional ill-conditioned inverse problem, has been greatly revitalized by the recent development of convolutional neural networks (CNN). These CNN-based methods generally map a low-resolution image to its corresponding high-resolution version with sophisticated network structures and loss functions, showing impressive performances. This paper provides a new insight on conventional SISR algorithm, and proposes a substantially different approach relying on the iterative optimization. A novel iterative super-resolution network (ISRN) is proposed on top of the iterative optimization. We first analyze the observation model of image SR problem, inspiring a feasible solution by mimicking and fusing each iteration in a more general and efficient manner. Considering the drawbacks of batch normalization, we propose a feature normalization (F-Norm, FN) method to regulate the features in network. Furthermore, a novel block with FN is developed to improve the network representation, termed as FNB. Residual-in-residual structure is proposed to form a very deep network, which groups FNBs with a long skip connection for better information delivery and stabling the training phase. Extensive experimental results on testing benchmarks with bicubic (BI) degradation show our ISRN can not only recover more structural information, but also achieve competitive or better PSNR/SSIM results with much fewer parameters compared to other works. Besides BI, we simulate the real-world degradation with blur-downscale (BD) and downscale-noise (DN). ISRN and its extension ISRN+ both achieve better performance than others with BD and DN degradation models., This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

This paper proposes a multi-data driven hybrid learning method for weekly photovoltaic (PV) power scenario forecast that is coordinately driven by weather forecasts and historical PV power output data. Patterns of historical data and weather forecast information are simultaneously captured to ensure the quality of the generated scenarios. By combining bicubic interpolation and bidirectional long-short term memory (BiLSTM), a super resolution algorithm is first presented to enhance the time resolution of weather forecast data from three hours to one hour and increase the precision of weather forecasting. A weather process-based weekly PV power classification strategy is proposed to capture the coupling relationships between meteorological elements, continuous weather changes and weekly PV power. A gated recurrent unit (GRU)-convolutional neural network (CNN)-based scenario forecast method is developed to generate weekly PV power scenarios. Evaluation indices are presented to comprehensively assess the quality of the generated weekly scenarios of PV power. Finally, the PV power, weather observation and weather forecast data collected from five PV plants located in Northeast Asia are used to verify the effectiveness and correctness of the proposed method.

Recent years have witnessed the remarkable success of deep learning-based single image super-resolution (SISR) methods. However, most of the existing SISR methods assume that low-resolution (LR) images are purely bicubic downsampled from high-resolution (HR) images. Once the actual degradation is not bicubic, their outstanding performance is hard to maintain. Since the real-world image degradation process can be modeled by a combination of downsampling, blurring, and noise, several SR methods have been proposed to super-resolve LR images with multiple blur kernels and noise levels. However, these SR methods require prior knowledge of the degradation process, which is difficult to obtain in practical applications. To address these issues, we propose a degradation reconstruction loss (DRL), which captures the degradation-wise differences between SR images and HR images via a degradation simulator. Empowered by the degradation simulator, the proposed loss, and an efficient SR network, a blind SR framework (SRDRL) without prior knowledge that can handle multiple degradations is formed. Extensive experimental results demonstrate that the proposed SRDRL outperforms the state-of-the-art blind SR methods and denosing+SR methods on multi-degraded datasets. The degradation reconstruction loss can be a plug-and-play loss for existing SR methods to handle multiple degradations. The source code can be found at https://github.com/FVL2020/SRDRL.

Photoacoustic microscopy (PAM) leverages the optical absorption contrast of blood hemoglobin for high-resolution, multi-parametric imaging of the microvasculature in vivo. However, to quantify the blood flow speed, dense spatial sampling is required to assess blood flow-induced loss of correlation of sequentially acquired A-line signals, resulting in increased laser pulse repetition rate and consequently optical fluence. To address this issue, we have developed a sparse modeling approach for blood flow quantification based on downsampled PAM data. Evaluation of its performance both in vitro and in vivo shows that this sparse modeling method can accurately recover the substantially downsampled data (up to 8 times) for correlation-based blood flow analysis, with a relative error of 12.7 ± 6.1 % across 10 datasets in vitro and 12.7 ± 12.1 % in vivo for data downsampled 8 times. Reconstruction with the proposed method is on par with recovery using compressive sensing, which exhibits an error of 12.0 ± 7.9 % in vitro and 33.86 ± 26.18 % in vivo for data downsampled 8 times. Both methods outperform bicubic interpolation, which shows an error of 15.95 ± 9.85 % in vitro and 110.7 ± 87.1 % in vivo for data downsampled 8 times.

Recent studies have shown that deep-learning (DL)-based methods lead to improved performance in video satellite super-resolution (SR). However, the vast majority of prior work is supervised, which is restricted to artificially generated training data (e.g., predetermined bicubic downsampling). Unfortunately, in the real world, the low-resolution (LR) satellite video frames rarely obey these restrictions. To solve this problem, we resort to unsupervised learning and propose a video satellite SR method by using only a single video. The single video SR (SingleVSR) method takes advantage of the power of DL without relying on prior high-resolution (HR) and LR pairs. In the training phase, the LR frames are alternately processed by both downsampling network (i.e., NetLR) and upsampling network (i.e., NetHR). The losses obtained by LR frames and network outputs are used to optimize both NetLR and NetHR. In the testing phase, the trained NetHR is applied to generate the SR results of LR frames. In contrast to the existing video satellite SR methods, our SingleVSR does not require any assumption on degradation or any additional training data except for the single video to be tested. Experiments performed on Jilin-1 and OVS-1 satellite videos demonstrate the superiority of the proposed method.

As the solutions of partial differential equations (PDEs), PDE surfaces provide an effective way for physical-based surface design in surface modeling. The bicubic B-spline surface is a useful tool for surface modeling in computer aided geometric design (CAGD). In this paper, we present a method for generating bicubic B-spline surfaces with the uniform knots and the quasi-uniform knots from the sixth order PDEs. From the given boundary condition, based on the cubic B-spline basis representation and the PDE mask, the resulting bicubic B-spline surface can be generated uniquely. The boundary condition is more flexible and can be applied for curvature-continuous surface design, surface blending and hole filling. Some representative examples show the effectiveness of the presented method.

This work was aimed to study the analgesic effect of pudendal nerve block on obstetrics and gynecology under the guidance of ultrasound image based on optimized fast super resolution reconstructed convolutional neural network (FSRCNN) algorithm. A total of 110 primiparas from hospital who gave birth through vagina were randomly rolled into experimental group (55 cases) and control group (55 cases). The optimized FSRCNN algorithm was constructed, compared with the FSRCNN algorithm and the Bicubic algorithm and applied to 110 cases of maternal patients undergoing perineotomy under ultrasound image-guided pudendal nerve block. Visual analogue scoring (VAS), incision suture pain VAS score, occurrence of complications, puerpera labor time, and newborn weight were recorded and compared, so did Apgar score of newborns, numbness of maternal thigh, recovery of puncture site, and satisfaction of maternal analgesia. The results showed that the peak signal-to-noise ratio (PSNR) of the high-resolution image reconstructed by the FSRCNN algorithm was 32.68 dB and that reconstructed by the optimized FSRCNN algorithm was 32.19 dB. The PSNR of the Bicubic algorithm processed image was 28.51 dB. In the lateral resection of episiotomy in the second stage of labor, the visual analog score (2.3 ± 1.5 points) of the experimental group was inferior to that of the control group (7.1 ± 2.6 points) ( P < 0.05 ). The visual analogue score of stitch pain (1.3 ± 0.8 points) was also inferior to that of the control group (5.2 ± 1.9 points) ( P < 0.05 ). Moreover, the satisfaction of the parturients in the experimental group (9.86 ± 0.41 points) was considerably superior to that of the control group (7.36 ± 1.25 points) ( P < 0.05 ). In short, the optimized FSRCNN algorithm had a short training time and good reconstruction effect. Ultrasound-guided pudendal nerve block had a substantial analgesic effect on the second stage of labor and improved parturients’ satisfaction.

Convolutional neural networks (CNNs) based super-resolution (SR) methods have achieved remarkable progress in recent years. Most of these methods assume that the degradation is known and fixed, such as bicubic downsampling. However, the performance of CNNs-based methods suffers from a severe drop when the actual degradation mismatch the training one. This paper proposes a prior correction network (PCNet) to solve the blind SR problem, which makes CNNs-based super-resolvers trained in a fixed blur kernel but applied to other unknown blur kernels. The PCNet consists of a kernel estimation network, a correction filter module, and a correction refinement network. The kernel estimation network aims to estimate unknown blur kernel from the input low-resolution (LR) image. The correction filter module then transfers the estimated degraded domains to adapt to specific degradations (e.g., bicubic downsampling). Finally, the correction refinement network adjusts the corrected LR image to eliminate the influence of blur kernel mismatch or misestimate. Experimental results on diverse datasets show that the proposed PCNet, combined with existing CNNs-based SR methods, outperforms other state-of-the-art algorithms for blind SR. Code is available at: \url{ https://github.com/caoxiang104/PCNet }.

BACKGROUND Fetal cardiac magnetic resonance imaging (MRI) improves the diagnosis of congenital heart defects, but is sensitive to fetal motion due to long image acquisition time. This may be overcome with faster image acquisition with low resolution, followed by image enhancement to provide clinically useful images. PURPOSE To combine phase-encoding undersampling with super-resolution neural networks to achieve high-resolution fetal cine cardiac MR images with short acquisition time. STUDY TYPE Prospective. SUBJECTS Twenty-eight fetuses (gestational week 36 [interquartile range 33-38 weeks]). FIELD STRENGTH/SEQUENCE 1.5 T, balanced steady-state free precession (bSSFP) cine sequence. ASSESSMENT Images were acquired using fully sampled Doppler ultrasound-gated clinical bSSFP cine as reference, with equivalent cine sequences with decreased phase-encoding resolution (25%, 33%, and 50% of clinical standard). Two super-resolution methods based on convolutional neural networks were proposed and evaluated (phasrGAN and phasrresnet). Data were partitioned into training (36 cine slices), validation (3 cine slices), and test sets (67 cine slices) without overlap. Conventional reconstruction methods using bicubic interpolation and k-space zeropadding were used for comparison. Three blinded observers scored image quality between 1 and 10. STATISTICAL TESTS Image scores are reported as median [interquartile range] and were compared using Mann-Whitney's nonparametric test with P

Visible light communications (VLC) based indoor positioning is a promising approach to serve an increasing need for location-aware services. Received signal strength (RSS) fingerprints based visible light positioning (VLP) was shown to achieve highly accurate, yet simple, VLP systems. It needs, however, large training datasets; whose collection is labor intensive. In this work, an artificial dataset generation methodology based on a modified pathloss model is proposed, where the pathloss exponent is assumed to be variable. Four interpolation techniques were investigated for the estimation of the variable pathloss exponent, where bicubic interpolation showed the best results in the simulation using a uniform sparse grid. The method is specially conceived to address the site surveying demand while attaining high positioning accuracy under non-line-of-sight (NLOS) conditions with sparse offline measurements. The technique is not limited to visible light applications since it depends only on the RSS and the pathloss model. Moreover, by using the weighted k -nearest-neighbor ( k -NN) machine learning technique, this work designs an accurate VLP system based on sparse fingerprints. Simulation results show an average positioning error of 2.48 cm using weighted k -NN (W k -NN) with only 53 offline measurements in a 25 m $^2$ area. Furthermore, experimental results show an average positioning error of 3.04 cm with only 18 offline measurements in an area of approximately 9 m $^2$ , and an average positioning error of 1.92 cm using only 16 offline measurements in an area of 6 m $^2$ .

Aero‐engine blades crack detection is one of the important tasks in daily ground maintenance, crack is a kind of texture feature, due to the random distribution, irregular shape and vague characteristics, which is still a challenging task to realize automatic detection in working environment. A detection model based on the Yolov4‐tiny is proposed that is universal and focuses more on the characteristics of cracks, and it is implemented in embedded device. First, in order to distinguish the cracks and noises, an improved attention module is introduced into the backbone of Yolov4‐tiny to enhance the model's capability to focus on crack areas; second, in order to improve the effect of multi‐scale feature fusion, the bicubic interpolation is implemented in upsampling module; finally, in order to solve the redundant detection results of bounding‐boxes in crack areas, the optimized non‐maximum suppression method is proposed to make the detection results better corresponding to the groundTruth. The robustness of proposed detection model was demonstrated by evaluating varying lighting and noise images. The average precision on integrated datasets is 81.6%, which outperforms the original Yolov4‐tiny by an increase of 12.3%.

Current research in iris recognition is moving towards enabling more relaxed acquisition conditions. This has effects on the quality of acquired images, with low resolution being a predominant issue. Here, we evaluate a super-resolution algorithm used to reconstruct iris images based on Eigen-transformation of local image patches. Each patch is reconstructed separately, allowing better quality of enhanced images by preserving local information. Contrast enhancement is used to improve the reconstruction quality, while matcher fusion has been adopted to improve iris recognition performance. We validate the system using a database of 1,872 near-infrared iris images. The presented approach is superior to bilinear or bicubic interpolation, especially at lower resolutions, and the fusion of the two systems pushes the EER to below 5% for down-sampling factors up to a image size of only 13x13., Published at Intl Conf on Biometrics: Theory, Apps and Systems, BTAS 2016

The text in a low-resolution (LR) image is usually hard to read. Super-resolution (SR) is an intuitive solution to this issue. Existing single image super-resolution (SISR) models are mainly trained on synthetic datasets whose LR images are obtained by performing bicubic interpolation or gaussian blur on high-resolution (HR) images. However, these models can hardly generalize to practical scenarios because real-world LR images are more difficult to super-resolve. The newly proposed TextZoom dataset is the first dataset for real-world text image super-resolution. We propose a new model termed TSRGAN trained on this dataset. First, a discriminator is designed to prevent the SR network from generating over-smoothed images. Second, we introduce triplet attention into the SR network for better representational ability. Moreover, besides L 2 loss and adversarial loss, wavelet loss is incorporated to help reconstruct sharper character edges. Since TextZoom provides text labels, the recognition accuracy of scene text recognition (STR) model can be used to evaluate the quality of SR images. It can reflect the performance of text image SR models better than traditional SR evaluation metrics such as PSNR and SSIM. Comprehensive experiments show the superiority of our TSRGAN. Compared with the state-of-the-art method, the proposed TSRGAN improves the average recognition accuracy of ASTER, MORAN and CRNN by 0.8%, 1.5% and 3.2% on TextZoom respectively.

An adaptive directional cubic convolution interpolation method for integrated circuit (IC) chip defect images is proposed in this paper, to meet the challenge of preserving edge and texture information. In the proposed method, Otsu thresholding technique is employed to distinguish strong edge pixels from weak ones and texture regions, and estimate the direction of strong edges, adaptively. Boundary pixels are pre-interpolated using the original bicubic interpolation method to help improve the interpolation accuracy of the interior pixels. The experimental results of both classic test images and IC chip defect images demonstrate that the proposed method outperforms the competing methods with better edge and texture preservation, interpolation quality, more natural visual effect of the interpolated images and reasonable computational time. The proposed method can provide high quality IC chip images for defect detection and has been successfully applied on practical vision inspection for IC chips

The deep neural network has achieved good results in medical image superresolution. However, due to the medical equipment limitations and the complexity of the human body structure, it is difficult to reconstruct clear cardiac magnetic resonance (CMR) superresolution images. To reconstruct clearer CMR images, we propose a CMR image superresolution (SR) algorithm based on multichannel residual attention networks (MCRN), which uses the idea of residual learning to alleviate the difficulty of training and fully explore the feature information of the image and uses the back-projection learning mechanism to learn the interdependence between high-resolution images and low-resolution images. Furthermore, the MCRN model introduces an attention mechanism to dynamically allocate each feature map with different attention resources to discover more high-frequency information and learn the dependency between each channel of the feature map. Extensive benchmark evaluation shows that compared with state-of-the-art image SR methods, our MCRN algorithm not only improves the objective index significantly but also provides richer texture information for the reconstructed CMR images, and our MCRN algorithm is better than the Bicubic algorithm in evaluating the information entropy and average gradient of the reconstructed image quality.

This paper mainly investigates the passive realization problems of bicubic (third-order) impedances as damper-spring-inerter networks consisting of no more than five elements. First, the special case where a bicubic impedance contains a pole or a zero on the imaginary axis or at infinity is discussed. Then, assuming that there is no pole or zero on the imaginary axis or at infinity, the realizations of bicubic impedances as five-element networks are investigated. Necessary and sufficient conditions for the realizability as five-element series-parallel networks and as five-element non-series-parallel networks are derived, respectively, where 22 series-parallel configurations and 11 non-series-parallel configurations are presented to cover the conditions. Finally, two numerical examples together with positive-real controller designs for a quarter-car suspension system are presented for illustrations. The results of this paper can contribute to the synthesis of low-complexity passive mechanical (or electrical) networks, which are motivated by the synthesis and design of inerter-based vibration control systems.

The resolution of ocean reanalysis datasets is generally low because of the limited resolution of their associated numerical models. Low-resolution ocean reanalysis datasets are therefore usually interpolated to provide an initial or boundary field for higher-resolution regional ocean models. However, traditional interpolation methods (nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation) lack physical constraints and can generate significant errors at land-sea boundaries and around islands. In this paper, a machine learning method is used to design an interpolation algorithm based on Gaussian process regression. The method uses a multiscale kernel function to process two-dimensional space meteorological ocean processes and introduces multiscale physical feature information (sea surface wind stress, sea surface heat flux, and ocean current velocity). This greatly improves the spatial resolution of ocean features and the interpolation accuracy. The effectiveness of the algorithm was validated through interpolation experiments relating to sea surface temperature (SST). The root mean square error (RMSE) of the interpolation algorithm was 38.9%, 43.7%, and 62.4% lower than that of bilinear interpolation, bicubic interpolation, and nearest neighbor interpolation, respectively. The interpolation accuracy was also significantly better in offshore area and around islands. The algorithm has an acceptable runtime cost and good temporal and spatial generalizability.

Medical research indicated that narrowing of the retinal blood vessels might be an early indicator of cardiovascular diseases; one of them is hypertensive retinopathy. This paper proposed the new staging method of hypertensive retinopathy by measure the ratio of diameter artery and vein (AVR). The dataset used in this research is the public Messidor color fundus image dataset. The proposed method consists of image resizing using bicubic interpolation, optic disk detection, a region of interest computation, vessel diameter measuring, AVR calculation, and grading the new categories of Hypertensive Retinopathy based on Keith-Wagener-Barker categories. The experiments show that the proposed method can determine the stage of hypertensive retinopathy into new categories.

Purpose Ultrasound (US) imaging has been widely used in diagnosis, image-guided intervention and therapy, where high-quality three-dimensional (3D) images are highly desired from sparsely acquired two-dimensional (2D) images. This study aims to develop a deep learning-based algorithm to reconstruct high-resolution 3D US images only reliant on the acquired sparsely distributed 2D images. Methods We propose a self-supervised learning framework using cycle consistent generative adversarial network (cycleGAN), where two independent cycleGAN models are respectively trained with paired original US images and two sets of low-resolution US images. The two sets of low-resolution US images are respectively obtained through down sampling the original US images along the two axes. In US imaging, in-plane spatial resolution is generally much higher than through-plane resolution. By learning the mapping from down-sampled in-plane low-resolution images to original high-resolution US images, cycleGAN can generate through-plane high-resolution images from original sparely distributed 2D images. Finally, high-resolution 3D US images are reconstructed by combining the generated 2D images from the two cycleGAN models. Results The proposed method was assessed on two different datasets. One is automatic breast ultrasound (ABUS) images from 70 breast cancer patients, the other is collected from 45 prostate cancer patients. By applying a spatial resolution enhancement factor of 3 to the breast cases, our proposed method achieved the mean absolute error (MAE) value of 0.90±0.15, the peak signal-to-noise ratio (PSNR) value of 37.88±0.88 dB, and the visual information fidelity (VIF) value of 0.69±0.01, which significantly outperforms bicubic interpolation. Similar performances have been achieved using the enhancement factor of 5 in these breast cases and using the enhancement factor of 5 and 10 in the prostate cases. Conclusions We have proposed and investigated a new deep learning-based algorithm for reconstructing high-resolution 3D US images from sparely acquired 2D images. Significant improvement on through-plane resolution has been achieved by only using the acquired 2D images without any external atlas images. Its self-supervision capability could accelerate high-resolution US imaging.

For satellite image processing using deep learning, the images are first resized to one standard size and then applied to the convolutional neural network(CNN) model to train and test the model. The primary step of preprocessing the images is resizing, which is done by various approaches such as bicubic interpolation, bilinear interpolation, and nearest neighbour interpolation. The impact of resizing technique on the performance of the deep learning model is unexplored yet. In the proposed work, the authors have proposed a CNN architecture and investigated the impact of resizing technique on the performance of trained deep learning model and evaluating the uncertainty of classification. The classification performed by the proposed model is a multilabel classification problem with 6 possible labels denoting land coverage. The proposed model is trained using SAT-6 dataset images. The SAT-6 dataset contains images of size $$28 \times 28$$ , which are resized to $$56 \times 56$$ using nearest neighbor interpolation for the training of the model, and testing is done using the aforementioned three interpolation techniques. Finally, the model performance is evaluated in term of specificity, sensitivity, accuracy, and uncertainty of the classification. This paper’s contribution is twofold – first, to put forward a generic model that will help other remote-sensing applications, and second to evaluate the uncertainty of the classification task by varying the image resizing technique.

Gas-liquid slug flow in horizontal pipes occurs frequently in industrial applications and is harmful to the piping systems owing to its intermittent structures. Therefore, it is essential to accurately describe the flow structure and investigate the flow mechanism. A non-intrusive fluorescence imaging method is employed to quantitatively reconstruct the slug bubble in a 15mm ID horizontal pipe in this article. Optical distortion correction method is firstly developed to improve the measurement accuracy of the instantaneous gas holdup, and then the time-average axial bubble velocity is simultaneously measured according to the feature matching of cross-sectional images. Additionally, the influence of phase superficial velocity on the interface stability is investigated through the phase holdup and the circumferential curvature distribution analysis. The typical three-dimensional structures of slug bubble are finally reconstructed through the velocity-based bicubic interpolation and median fusion algorithm. Previous correlations and reference instrumentation are used to evaluate the proposed method. Comparison results show that the method can effectively measure the flow parameters and reconstruct slug bubbles.

Very high-resolution satellite imagery from the latest generation commercial platforms provides an unprecedented capacity for imaging the Earth with very high spatial detail. However, these data are generally expensive, particularly if large areas or temporal sequences are required. In recent years, lower quality imagery has been enabled through the launch of constellations of small satellites with short revisit time. In this article, we apply for the first time a statistical approach to downscale and bias-correct these multispectral satellite data using the information contained in a limited training set of very high-resolution images. The technique, based on the direct sampling algorithm, aims at extending the coverage of high-resolution images by sampling data from a training data set, where similar lower resolution data patterns are found. Unlike the majority of the current downscaling techniques, the approach does not require colocated fine-resolution data, but it is based on the use of training images similar to the target zone. A novel specific setup is proposed, which is adaptive to different types of landscapes with no additional user effort. The results show that the proposed technique can generate more realistic images than the traditional approaches based on the parametric bias correction and bicubic interpolation. In particular, properties such as the intensity histogram, spatial correlation, and connectivity are accurately preserved. The proposed approach can be used to extend the footprint of the high-resolution images to generate new time frames or to downscale the remote sensing imagery based on a distant but structurally similar training image.

The purpose of this paper is to present an appropriate numerical method for solving linear two-dimensional weakly singular stochastic integral equations. This method is based on bicubic B-spline interpolation, Gauss–Legendre quadrature formula, and two-dimensional Ito approximation. The proposed method is used to transform the solution of mentioned problem to the linear system of algebraic equations which can be solved by using an appropriate method. Convergence analysis of the approach has been investigated. In the end, two examples are given to show efficiency and accuracy of the method and their numerical results are reported.

Bicubic interpolation is widely used in real-time image processing systems because of its quality. The real-time implementation of bicubic interpolation requires a lot of hardware resources, especially the number of multipliers because it represents high computational complexity. In this article, a set of algorithms that approximate the bicubic interpolation and reduce the hardware resource consumption are proposed. The proposed algorithms are based on combining linear and cubic interpolations. These algorithms are surveyed and compared in terms of interpolation quality, number of adders, number of multipliers, adaptive logic modules, lookup tables (LUTs), registers, and maximum operating frequency. These algorithms are implemented and tested on an Intel Cyclone V target. This article provides various choices of interpolation algorithms to cater to different application requirements, including accuracy, hardware resource consumption, and throughput performance. The implementation codes are available at github.com/DreamIP/Interpolation .

Aiming at insufficient detailed description problem caused by the loss of edges during a single low-resolution (LR) image’s reconstruction process, a novel algorithm for super resolution image reconstruction is proposed in this paper, which is based on fusion of internal structural self-similarity dictionary and external convolution neural network parameters learning model. Firstly, for solving training samples too scattered problem, besides external database, an internal database is constructed to learn a dictionary of the single image’s structural self-similarity by multi-scale decomposition approach. Secondly, nonlocal regularization constraint is calculated on the priori knowledge, which is obtained from the internal database of the single LR image. Thirdly, similar block pairs of high and low-resolution samples in the external database are input into a convolution neural network for learning the parameters of reconstructing model. After all, combined parameters learned and the internal dictionary, the single LR image is reconstructed, and by iterative back-projection algorithm its result is improved. Experimental results show that, compared with state-of-the-art algorithms, such as Bicubic, K-SVD algorithm and SRCNN algorithm, our method is more effective and efficient.

Advent of convolutional neural network (CNN) in the field of single image super-resolution (SISR) has shown immense improvement in the process of high resolution (HR) image generation. It involves an end-to-end mathematical mapping through non-linear feature extraction between low resolution (LR) and HR image. Performance of CNN can be improved by increasing depth of the architecture, which generally results in higher computational cost and running time. Also, the performance of CNN can be improved by providing more appropriate input. Presently, in most of the cases input image to a CNN is a LR image that is bi-cubically interpolated to the desired size of HR image. However, bicubic interpolation results into detail smoothing of the image. Therefore, in this paper, a hybrid of CNN and fractal interpolation based SISR algorithm is proposed for reconstruction of HR image. Here, a three layered light-weight CNN architecture is utilize which is capable of producing comparable performance with the traditional SISR techniques and fractal interpolation helps in better preservation of structural and textural properties of the HR image. Experimental results are provided to prove the efficacy of the algorithm proposed in the paper.

Traditionally, prior to compressing an RGB full-color image, for each converted $2\times 2$ CbCr block $B^{CbCr}$ , chroma subsampling only downsamples $B^{CbCr}$ , but without changing the luma block $B^{Y}$ at all. In the current research, a special linear interpolation-based, namely the COPY-based, chroma subsampling-first luma modification (CSFLM) study has attempted to change the luma block for enhancing the quality of the reconstructed RGB full-color image. In this paper, a fast and effective nonlinear interpolation, namely the bicubic convolution interpolation (BCI), based iterative luma modification method for CSFLM is proposed. In our iterative method, a BCI-based distortion function and its convex property proof are first provided. Next, based on the proposed convex distortion function, a pseudoinverse technique is applied to obtain the initial luma modification solution, and then an iterative method is proposed to improve the initial luma modification solution. Based on five testing image datasets, namely the IMAX, Kodak, SCI (screen content images), CI (classical images), and Video datasets, the thorough experimental results have demonstrated that on the newly released Versatile Video Coding (VVC) platform VTM-12.0, our iterative luma modification method achieves substantial quality, execution-time, and quality-bitrate tradeoff improvements when compared with the existing state-of-the-art methods.

High-resolution (HR) Mars images have great significance for studying the land-form features of Mars and analyzing the climate on Mars. Nowadays, the mainstream image super-resolution methods are based on deep learning or CNNs, which are better than traditional methods. However, these deep learning based methods obtain low-resolution(LR) images usually by using an ideal down-sampling method (e.g. bicubic interpolation). There are two limitations in the existing SR methods: 1) The paired LR-HR data by using such methods can achieve a satisfactory results when tested on an ideal datasets. But, these methods always fail in real Mars image super-resolution, since real Mars images rarely obey an ideal down-sampling rule. 2) The LR images obtained by ideal down-sampling methods have no noise while real Mars images usually have noise, which leads to the super-resolved images are not realistic in texture details. To solve the above-mentioned problems, in this article, we propose a novel two-step framework for Mars image super-resolution. Specifically, to address limitation 1), we focus on designing a new degradation framework by estimating blur-kernels. To address limitation 2), a Generative Adversarial Network (GAN) is trained to generate noise distribution. Extensive experiments on the Mars32k dataset demonstrate the effectiveness of the proposed method, and we achieve better qualitative and quantitative results compared to other SOTA methods.

Displays and content have various resolutions and aspect ratios, requiring an image downscaler to adaptively reduce the image resolution. However, research on downscaling has garnered less attention than upscaling, including super-resolution. In practical display systems, simple interpolation, such as a bicubic filter that cannot preserve image details well, is still widely used for image downscaling rather than frame optimization-based or learning-based methods because of following reasons: frame optimization-based methods can effectively preserve image details after downscaling but are difficult to implement due to hardware costs. Learning-based methods have not been developed because defining a target downscaled image for training is difficult and training all downscaling factors is impossible. We propose a novel kernel-learning-based image downscaler to improve detail-preservation quality while supporting arbitrary downscaling factors using simple linear mapping. For this, a method to produce the ideal target downscaling result considering aliasing artifacts and detail preservation after downscaling is proposed. Then, we propose a training technique using the positional relationship between input and output pixels and a hierarchical region analysis to reproduce target images through simple kernel-based linear mapping. Lastly, a kernel-sharing technique is proposed to generate downscaling results for downscaling factors using a minimum number of trained kernels. In the simulation results, the proposed method demonstrated excellent edge preservation by improving the recall, precision, and F1 score, measuring the edge consistency between input and downscaled images, by up to 0.141, 0.079, 0.053, respectively, compared to benchmark methods. In a paired-comparison-based user study, the proposed method obtained the highest preference among benchmark methods using simple operations.

This paper presents a spherical measure based spherical image representation(SMSIR) and sphere-based resampling methods for generating our representation. On this basis, a spherical wavelet transform is also proposed. We first propose a formal recursive definition of the spherical triangle elements of SMSIR and a dyadic index scheme. The index scheme, which supports global random access and needs not to be pre-computed and stored, can efficiently index the elements of SMSIR like planar images. Two resampling methods to generate SMSIR from the most commonly used ERP(Equirectangular Projection) representation are presented. Notably, the spherical measure based resampling, which exploits the mapping between the spherical and the parameter domain, achieves higher computational efficiency than the spherical RBF(Radial Basis Function) based resampling. Finally, we design high-pass and low-pass filters with lifting schemes based on the dyadic index to further verify the efficiency of our index and deal with the spherical isotropy. It provides novel Multi-Resolution Analysis(MRA) for spherical images. Experiments on continuous synthetic spherical images indicate that our representation can recover the original image signals with higher accuracy than the ERP and CMP(Cubemap) representations at the same sampling rate. Besides, the resampling experiments on natural spherical images show that our resampling methods outperform the bilinear and bicubic interpolations concerning the subjective and objective quality. Particularly, as high as 2dB gain in terms of S-PSNR is achieved. Experiments also show that our spherical image transform can capture more geometric features of spherical images than traditional wavelet transform.

Background: High-resolution medical images often need to be downsampled because of the memory limitations of the hardware used for machine learning. Although various image interpolation methods are applicable to downsampling, the effect of data preprocessing on the learning performance of convolutional neural networks (CNNs) has not been fully investigated. Methods: In this study, five different pixel interpolation algorithms (nearest neighbor, bilinear, Hamming window, bicubic, and Lanczos interpolation) were used for image downsampling to investigate their effects on the prediction accuracy of a CNN. Chest X-ray images from the NIH public dataset were examined by downsampling 10 patterns. Results: The accuracy improved with a decreasing image size, and the best accuracy was achieved at 64 × 64 pixels. Among the interpolation methods, bicubic interpolation obtained the highest accuracy, followed by the Hamming window.

Image fusion is a visual enhancement technique that combines source images from different sensors to produce a more robust and informative fused image for subsequent processing or decision making. Infrared and visible light images share complementary properties that enable the production of robust and informative fused images. This paper proposed an infrared and visible image fusion method that improved the tetrolet framework to improve infrared and visible image fusion quality. First, the source image is enhanced by bicubic interpolation. The improved tetrolet transform then decomposes the enhanced source image; the high-frequency components are fused by convolutional sparse representation theory and combined with corresponding rules, and the low-frequency components are fused by defining ISER descriptors. Finally, we use the inverse transform to reconstruct the fused image. Qualitative and quantitative experimental results on five groups of typical infrared and visible image datasets demonstrate the proposed method's effectiveness. The proposed method exhibits better performances on subjective vision and objective indexes compared with the other state-of-the-art methods.

In this paper, a novel digital video resolution enhancement algorithm based on adaptive directional interpolation is proposed, where the directionality of the edge structure and the nonlocal self-similarity prior within the current frame as well as its adjacent frames are both considered. First, we establish the regularization equation that conforms to the prior model of a video frame and then take the classic bicubic interpolation result as the initial estimation to iteratively solve the restoration equation, in which the edge structures and contours in low resolution (LR) input are reconstructed to estimate and refine the desired high resolution (HR) output. Experimental results show that the proposed algorithm can effectively enhance the clarity of a video frame, with satisfying subjective visual quality and PSNR value.

[EN] In neurosciences, imaging techniques are used to diagnose diseases and assist in surgical interventions to determine the exact location and size of a tumor. International research has used different techniques for the fusion of Positron Emission Tomography and Computed Tomography because it allows us to observe metabolic functions in correlation with anatomical structures. From the study of these techniques and algorithms a software was developed that performs the fusion of neuroimaging, using the fusion scheme based on Wavelet. Bicubic interpolation was used to correct the images. As a discrete Wavelet transform, Haar's was used. An evaluation of the quality of the resulting images was performed, obtaining 512 × 512 × 24 bit resolution images and a signal-to-noise ratio with values greater than 18. It was verified by histograms that the merged image contains more information than the input images independent., [ES] En neurociencias se utilizan técnicas imagenológicas para el diagnóstico de enfermedades y la asistencia en intervenciones quirúrgicas para determinar la ubicación y dimensión exactas de un tumor. Investigaciones a nivel internacional han utilizado diferentes técnicas para la fusión de la Tomografía por Emisión de Positrones y la Tomografía Computarizada debido a que permite observar las funciones metabólicas en correlación con las estructuras anatómicas. A partir del estudio de estas técnicas y algoritmos se desarrolló un software que realiza la fusión de neuroimágenes, utilizando el esquema de fusión basado en Wavelet. Para el corregistro de las imágenes se utilizó la interpolación Bicúbica. Como transformada discreta de Wavelet se utilizó la de Haar. Se realizó una evaluación de la calidad de las imágenes resultantes, obteniendo imágenes de resolución 512 × 512 × 24 bits y una relación señal-ruido con valores superiores a 18. Se comprobó mediante histogramas que la imagen fusionada contiene más información que las imágenes de entrada independientes., Este trabajo ha sido realizado parcialmente gracias al apoyo del Centro de Informática Médica de la Universidad de las Ciencias Informáticas de la Habana y el Centro de Neurociencias de Cuba CNEURO, por el apoyo logístico y asesoría especializada en temas de Neuroinformática.

This paper studies the blind super-resolution of real low-quality and low-resolution (LR) images. Existing convolutional network (CNN) based approaches usually learn a single image super-resolution (SISR) model for a specific downsampler (e.g., bicubic downsampling, blurring followed by downsampling). The learned model, however, is tailored to the specific downsampler and fails to super-resolve real LR images which are degraded in more sophisticated and diverse manners. Moreover, the ground-truth high-resolution (HR) of real LR images are generally unavailable. Instead of learning from unpaired real LR-HR images or a specific downsampler, this paper learns blind SR network from a realistic, parametric degradation model by considering blurring, noise, downsampling, and even JPEG compression. In contrast to direct blind reconstruction of HR image, the proposed model adopts a cascaded architecture for noise estimation, blurring estimation, and non-blind SR, which can be jointly end-to-end learned from training data and benefit generalization ability. By taking the bicubicly upscaled LR image as input to non-blind SR, the proposed method can present a single unified model for blind SR with any upscaling factors and varying degradation parameters. Experimental results show that the proposed method performs favorably on synthetic and real LR images.

Convolutional neural networks (CNNs) have recently made impressive results for image super-resolution (SR). Our goal is to introduce a new image SR framework rely on a CNN. In this paper, the input image is decomposed into luminance channel and chromatic channels. A designed network based on a residual dense network is introduced to extract the hierarchical features from luminance part. The bicubic interpolation is simply used to upscale low resolution (LR) chromatic channels. However, this step degrades the chromatic channels. To tackle this issue, the SR reconstructed luminance channel is applied as the reference image in guided filters to promote the interpolated chromatic channels. Guided filters technique has ability to retain sharp edges and fine details from the reference image and carry them to the target images. Extensive experiments on several commonly used image SR testing datasets demonstrate that our framework has the ability to extract features and outperforms existing well-known techniques for image SR by LR image into the high resolution (HR) image efficiently.

To fill the development gap between computer-aided design (CAD) and computer-aided manufacturing (CAM) on non-uniform rational b-spline (NURBS), a general tool path planning method on NURBS surface is proposed in this study. The generated tool path is composed of a series of closed paths which are derived by mapping the contours of a shortest boundary geodesic map (SBGM) from parameter space to 3D Euclidean space. SBGM that is created in parameter space depicts the shortest boundary geodesic distance (SBG) from internal points on NURBS surface to surface boundaries. It is calculated in discrete form firstly and then reconstructed by using a bicubic smoothing spline algorithm to improve the smoothness of the generated tool path. The tool path interval is deduced according to the requirement of residual scallop height constraint and determined conveniently based on the SBG difference of the two adjacent paths. It is confirmed that the proposed tool path planning method can be applied on all kinds of NURBS surface without the requirements of any other post-processing procedures. Typical case studies and experiments are presented to verify the effectiveness of the proposed method. Comparing with other traditional approaches, the proposed tool path planning method has excellent performance in improving the smoothness and reducing the total length of tool path.

Increasing global water deficit and demand for yield improvement call for high-resolution monitoring of irrigation, crop water stress, and crops' general condition. To provide high spatial resolution with high-temperature accuracy, remote sensing is conducted at low altitudes using radiometric longwave thermal infrared cameras. However, the radiometric cameras' price, and the low altitude leading to low coverage in a given time, limit the use of radiometric aerial surveys for agricultural needs. This paper presents progress toward solving both limitations using algorithmic and computational imaging methods: stabilizing the readout of low-cost thermal cameras to obtain radiometric data, and improving the latter's low resolution by applying convolutional neural network-based super-resolution. The two methods were merged by an end-to-end algorithm pipeline, providing a large mosaicked image of the field. First, the potential capabilities of a joint estimation method to correct unknown offset and gain were simulated on remotely sensed agricultural data. Comparison to ground-truth measurements showed radiometric accuracy with a root mean square error (RMSE) of 1.3 °C to 1.8 °C. Then, the proposed super-resolution method was demonstrated on experimental and simulated remotely sensed agricultural data. Preliminary experimental results showed 50% improvement in image sharpness relative to bicubic interpolation. The performance of the algorithm was evaluated on 22 simulated cases at × 2 and × 4 magnification. Finally, image mosaicking using the proposed pipeline was demonstrated. A mosaicked image composed of sub-images pre-processed by the proposed computational methods resulted in a RMSE in temperature of 0.8 °C, as compared to 8.2 °C without the initial processing.

Magnetic resonance imaging (MRI) is a very effective method for identifying any abnormality in the structure and physiology of the spine. However, MRI is time consuming as well as costly. In this work, the authors propose an algorithm which can reduce the time of MRI and thus the cost, with minimal compromise on accuracy. They reconstruct a three-dimensional (3D) image of the spine from a sequence of 2D MRI slices along any one axis with reasonable slice gap. In order to preserve the image at the edges properly, they regenerate the 3D image by using a combination of bicubic and bilinear interpolation along the orthogonal axis. From the reconstructed 3D, they use a simple geometric method to slice out any possible location along any axis and get the information in that region. They have tested their algorithm on real data, and found that their algorithm reduces the time by 80%, with high internal data preservation accuracy of about 96%.

In medical image processing, the automatic analysis of pathology localization and the anatomical segmentation steps are more important. The Fundus images of Low resolution (LR) are not applicable to detect the retinal disease. The main aim of this paper is to enhance the resolution of the low-resolution retinal images obtained from the cheap imaging devices within less computational time and high accuracy. So, we proposed the fundus image with Super-Resolution and its performance via the Diagnostically Significant Area (DSA). This approach focuses only on the region of Interest (ROI) instead of concentrating on the entire image leading to less computational time by reducing the time complexity. Therefore, the Eigen MR inter-band feature, Energy MR intra-band feature, Shannon entropy and Sensitive Contrast Interest (SCI) are used to capture the clinical data from the selected region. Therefore, the DSA is determined by using Levenshtein based KNN classifier. Because of better classification outcomes, the Bicubic method is employed in the selected region to reduce the loss of reconstruction error. Experimentally, the implementation works are carried out in the platform of MATLAB with DRIVE and STARE database images are chosen. The super-resolution image performances are compared with different start of art techniques such as PSM, GR-SR, LLE, and SpC-SR. Finally, higher efficiency with low computational super-resolution fundus images is collected.

In this paper, we rst propose a fast and eective region-based depth map upsampling method, and then propose a joint upsampling and location map-free reversible data hiding method, simpled called the JUR method. In the proposed upsampling method, all the missing depth pixels are partitioned into three disjoint regions: the homogeneous, semi-homogeneous, and non- homogeneous regions. Then, we propose the depth copying, mean value, and bicubic interpolation approaches to reconstruct the three kinds of missing depth pixels quickly, respectively. In the proposed JUR method, without any location map overhead, using the neighboring ground truth depth pixels of each missing depth pixel, achieving substantial quality, and embedding capacity merits. The comprehensive experiments have been carried out to not only justify the execution-time and quality merits of the upsampled depth maps by our upsampling method relative to the state-of-the-art methods, but also justify the embedding capacity and quality merits of our JUR method when compared with the state-of-the-art methods.

Super-resolution reconstruction refers to the technique of reconstructing a high-resolution image from a single or a series of low-resolution images by digital image processing. This technology can not only increase the high-frequency information of the image, but also eliminate the low-resolution. Deep Learning has made breakthroughs in modern digital image processing. Compared to traditional algorithms, deep convolutional neural networks (DCNN) achieve superior performance on a series of challenging image-processing problems such as image classification and target detection. Enhancement Deep convolutional neural networks (EDCNN) learn through a large number of training samples, obtain relevant information within the image, and then use the information to achieve specific functions. EDCNN also has an excellent performance with remote sensing data. Performance evaluation was made with bicubic and other deep learning methods, EDCNN outperformed other deep learning algorithms.

In this study, an efficient self-learning method for image super-resolution (SR) is presented. In the proposed algorithm, the input image is divided into equal size patches. Using these patches, a dictionary is learned based on K-SVD, referred to as high resolution (HR) dictionary. Then, by down-sampling, the columns of the dictionary, called atoms, a low resolution (LR) version of the dictionary is obtained. An initial estimate of the SR image is constructed using the bicubic interpolation on the input image. Then in an iterative algorithm, the difference between the down-sampled version of the estimated SR image and the input image is obtained. This difference image, which includes reconstructed details is enlarged using sparse representation and LR/HR dictionaries. The enlarged detail is added to the latest reconstructed SR image. This process gradually improves the quality of the initial SR image. After several iterations, the reconstructed image is an SR version of the input image. Experimental results confirm that the proposed method performance is promising.

Recently, convolutional neural network (CNN) based models have shown great potential in the task of single image super-resolution (SISR). However, many state-of-the-art SISR solutions are reproducing some tricks proven effective in other vision tasks, such as pursuing a deeper model. In this paper, we propose a new solution (named as Multi-Receptive-Field Network - MRFN), which outperforms existing SISR solutions in three different aspects. First, from receptive field: a novel multi-receptive-field (MRF) module is proposed to extract and fuse features in different receptive fields from local to global. Integrating these hierarchical features can generate better mappings on recovering high-fidelity details at different scales. Second, from network architectures: both dense skip connections and deep supervision are utilized to combine features from the current MRF module and preceding ones for training more representative features. Moreover, a deconvolution layer is embedded at the end of the network to avoid artificial priors induced by numerical data pre-processing (e.g., bicubic stretching), and speed up the restoration process. Finally, from error modeling: different from $L1$ and $L2$ loss functions, we proposed a novel two-parameter training loss called Weighted Huber loss function which can adaptively adjust the value of back-propagated derivative according to the residual value, thus fit the reconstruction error more effectively. Extensive qualitative and quantitative evaluation results on benchmark datasets demonstrate that our proposed MRFN can achieve more accurate recovering results than most state-of-the-art methods with significantly less complexity.

The goal of computer vision is to identify objects of interest from images. Copy-move forgery is the process of copying one or more parts of an image and moved into another part of an equivalent image. The detection of copy-move images has many limitations in recognition of objects. In this paper, the proposed work uses Speeded Up Robust Feature (SURF) feature extraction, and the specific object is recognized with the help of the support vector machine. When copy-move forgery was performed, some modifications were done to the image. For instance, turning, scaling, darkening, compression, and noise addition are applied to make effective impersonation forgeries. Here, feature matching process uses the image rotate function, which consists of bicubic and crop operations, and calculates the difference using the blend, scale and joint operation. The results show that forged images are extracted from a given set of test images. The test results exhibit that the proposed technique can get noteworthy and impressive results.

Thermal distribution and fluctuation in any piping component due to turbulent mixing of flows with different temperatures vary greatly. Usually, computational fluid dynamics (CFD) tools are used for estimation of flows in piping components. Fatigue that results from fluctuating thermal mass flow across the components can be computed by coupling the CFD results with structural mechanics based finite element (FE) results. However, this procedure is laborious and computationally very expensive. A fluid temperature function has been developed in this paper as a function of internal wall coordinates and time by interpolating experimental or CFD results. Bicubic interpolation function has been used for accurate interpolation. Finally, a thermal transient FE analysis for an actual Tee from a nuclear power plant (NPP) was performed by using the developed fluid temperature function and interpolated CFD results.

Multi-focus image fusion methods combine two or more images which have blurred and defocused parts to create an all-in-focused image. All-in-focused image has more information, clearer parts and clearer edges than the source images. In this paper, a new approach for multi-focus image fusion is proposed. Firstly, the information of source images is enhanced using bicubic interpolation-based super-resolution method. Secondly, source images with high resolution are decomposed into four sub-bands which are LL (low-low), LH (low-high), HL (high-low) and HH (high-high) using Stationary Wavelet Transform with dmey (Discrete Meyer) filter. Then, a new fusion rule which depend on gradient-based method with sobel operator is implemented to create fused images with good visuality. The weight coefficients which show the importance rates of corresponding pixels in source images for fused image are calculated using designed formula based on gradient magnitudes. The each pixel of fused sub-bands is created using these weight coefficients and fused image is reconstructed using Inverse Stationary Wavelet Transform. Lastly, the performance evaluation of proposed method is measured using three different metrics which are objective, subjective and time criterion metrics. Besides these features, the new dataset which is different from the datasets in the literature is created and used firstly in this paper. The results show that the proposed method produces high quality images with clear edges and transmits most of the information of source images into all-in-focused image.