Image Quality Evaluation Method Research Articles

Restoration of Laser Interference Image Based on Large Scale Deep Learning

Although the deep learning method has achieved some results in the challenging task of repairing the missing areas of images over the years, there is no report on the images disturbed by the laser entering the field of view. We propose a restoration model that can restore the laser interference image, the corresponding countermeasure network deep learning model and a new model training method, and no additional manual marking work is required in the data training set. After a large number of experiments, the loss function can rapidly converge under this method, accurately give the reasonable reconstruction results of the image disturbed by laser, and significantly improve the scores of many common image quality evaluation methods: high-quality repair results are obtained on the laser interference composite dataset of face (CelebA), Stanford automobile, aircraft, buildings (Facade) and satellite images. It has the characteristics of fast training speed, strong robustness, modular model design and wide application.

DVL2021: An ultra high definition video dataset for perceptual quality study

This paper describes an ultra high definition (UHD) video dataset named DVL2021 for the perceptual study of video quality assessment (VQA). To our knowledge, DVL2021 is the first authentically distorted 4K (3840 × 2160) UHD video quality dataset. The dataset contains 206 versatile 4K UHD video sequences, which are all collected in in-the-wild scenarios. Each sequence is captured at 50 frames per second (fps), stored in raw 10-bit 4:2:0 YUV format, and has a duration of 10 s. Following the subjective evaluation method of TV image quality granted by ITU-R BT.500-13, 32 unique participants take part in the manual annotation process, whose ages are from teenage to sixties (32.7 years old on average). DVL2021 has the following merits: (1) enormous variety of video contents, (2) captured by different types of cameras, (3) complex types and multiple levels of authentic distortion, (4) broadly distributed temporal/spatial information, and (5) a wide spectrum of mean opinion scores (MOS) distribution. Furthermore, we conduct a benchmark experiment by evaluating several mainstream VQA methods on DVL2021. The baseline results are higher than 0.75 in Spearman’s rank order correlation coefficient (SROCC) metric. Our study provides a basis for the UHD VQA problem. DVL2021 is publicly available at https://github.com/GZHU-DVL/DVL2021.

Image Quality Evaluation of Sanda Sports Video Based on BP Neural Network Perception.

In the special sports camera, there are subframes. A lens is composed of multiple frames. It will be unclear if a frame is cut out. The definition of video screenshots lies in the quality of video. To get clear screenshots, we need to find clear video. The purpose of this paper is to analyze and evaluate the quality of sports video images. Through the semantic analysis and program design of video using computer language, the video images are matched with the data model constructed by research, and the real-time analysis of sports video images is formed, so as to achieve the real-time analysis effect of sports techniques and tactics. In view of the defects of rough image segmentation and high spatial distortion rate in current sports video image evaluation methods, this paper proposes a sports video image evaluation method based on BP neural network perception. The results show that the optimized algorithm can overcome the slow convergence of weights of traditional algorithm and the oscillation in error convergence of variable step size algorithm. The optimized algorithm will significantly reduce the learning error of neural network and the overall error of network quality classification and greatly improve the accuracy of evaluation. Sanda motion video image quality evaluation method based on BP (back propagation) neural network perception has high spatial accuracy, good noise iteration performance, and low spatial distortion rate, so it can accurately evaluate Sanda motion video image quality.

Blind quality assessment of night-time image

High-quality night-time imaging is crucial to video surveillance, automatic drive and consumer electronics. However, different from day-time imaging, night-time imaging suffers from some disadvantages, such as low light, uneven illumination, difficult focusing, etc., which raises a great concern to the night-time imaging quality. Accordingly, a practical night-time image quality evaluation method is very promising to control and improve the night-time imaging system. Toward this end, in this paper, we propose a blind image quality assessment (BIQA) method to quantify the night-time image quality. Specifically, in the proposed method, we measure the night-time image quality by investigating the fundamental image properties, which are highly relevant to the image quality, such as the brightness, saturation, sharpness, noiseness, contrast and the semantics. Specific features are designed to characterize the image properties properly. Then we employ the support vector regression (SVR) method to infer the image quality with the extracted quality-aware features. The proposed BIQA method for night-time images is thoroughly evaluated on a representative night-time image database. Experimental results demonstrate that the proposed BIQA method for night-time images achieves superior prediction performance to other state-of-the-art BIQA methods.

Remote sensing image quality evaluation based on deep learning

Remote sensing technology is an effective tool for sensing the earth’s surface. With the continuous improvement of remote sensing technology, remote sensing detectors can obtain more spectral and spatial information, including clear feature contours, complex texture features and spatial layout rules. This information was detected in mineral resources. Surface substance identification, water pollution information monitoring and many other aspects have played an important role. The coding algorithm and defects, storage algorithm and interference from atmospheric cloud radiation information during the imaging process lead to varying degrees of distortion and deterioration of remote sensing images during imaging, transmission and storage. This makes it difficult to process, analyze and apply remote sensing images. Therefore, the design of a reasonable remote sensing image quality evaluation method is not only conducive to the remote sensing image quality evaluation in the real-time processing system of remote sensing image, but also conducive to the optimization of remote sensing image system and image processing algorithm. The application is worthwhile. In this paper, the deteriorating features of remote sensing images will change the statistical distribution. We propose a method for evaluating the quality of remote sensing images in depth learning. Feature learning and blurring as well as noise intensity classification for image remote sensing using convolutional neural network are carried out. The evaluation model is modified by masking effect and perceptual weighting factor, and the quality evaluation results of remote sensing images are obtained according to human vision. The research shows that this method can effectively solve the problem of removing and evaluating the noise of remote sensing image, and can effectively and accurately evaluate the quality of remote sensing image. It is also consistent with subjective assessment and human perception.

Overview and research progress of no reference image quality evaluation methods

Image quality will greatly affect the acquisition of human visual information, but the image will inevitably cause a certain degree of distortion in the process of transmission, compression, coding and so on. At this time, the accuracy of image quality evaluation (IQA) is particularly important, which also makes IQA an important research direction in the field of image processing. IQA methods are divided into subjective image quality evaluation method and objective image quality evaluation method.there are many subjective evaluation methods, and the international telecommunications union (ITU) has also proposed many subjective evaluation methods and standards. there are three most commonly used methods, namely, the double stimulation continuous quality classification method, the single stimulation continuous quality classification method and the double stimulation damage classification method. Correspondingly, there are three kinds of objective evaluation methods, full reference, half reference and no reference. At present, this paper selects the research focus of IQA field-no reference image quality evaluation method as the main research object, summarizes the advantages and disadvantages of each non-reference evaluation method, and prospects the future development of the non-reference image quality evaluation method.

A structured light image quality evaluation method based on no-reference quality assessment

Affected by noise, color, and light blocking, there are some large error data contained in the results of structured light 3D measurement. In the absence of effective criterion, the data can only be recognized artificially. In this paper, a criterion based on Human Vision System (HVS), which can be applied to recognize the large error data by evaluating the noise variance level of the stripe images, and a method used to improve the level of measurement accuracy are proposed. According to the proposed method, the no-reference quality assessment is used to evaluate the quality of stripe images. With the noise evaluation coefficient, the large error data can be identified automatically. The experimental results show that this method is efficient in detecting large error data in the stripe images.

Fringe Pattern Denoising Using Two-Dimensional Variational Mode Decomposition (2D-VMD) Method for Inspection of Flatness of Reduced Surfaces

The flatness is a major element for any functional precision surface. It becomes extremely important in the miniaturization of the systems, which require the components of reduced sizes. There is a need to develop appropriate methods to solve measurement problems in industrial design, manufacturing of production and process monitoring. In this work, we used an optical method based on the moire interferometry for topography in full-field and to control the flatness of the reduced surfaces having a size of 30 mm2 and more. This technique makes it possible to palpate optically the surface and to give an instantaneous topography of surface in real-time. We have developed an optical set-up which it is possible to inspect in a nondestructive way and without contact, the surface quality. Unfortunately, the moire fringes taken by camera contain excessive noise. That is why we proposed to use the Two-Dimensional Variational Mode Decomposition (2D-VMD) method to reduce the random noise from the fringe patterns and to ameliorate the profiles and the residual images from which we can control the presence of defects flatness of reduced surfaces. We have also used three objective image quality evaluation methods (PSNR, SNR and correlation coefficient) to compare between the denoising results given by 2D-VMD with those obtained by the Bidimensional Empirical Mode Decomposition (BEMD).

Design and optimization method of a convex blazed grating in the Offner imaging spectrometer

The convex reflective diffraction grating is an essential optical component in Offner systems, which has been widely used in imaging spectrometers. We propose a new design and optimization method for the convex blazed grating in the Offner imaging spectrometer. The method integrates the macro- and microdesign of the optical system, and it can be used to design and optimize the convex blazed grating with high diffraction efficiency. Traditional geometric optics theory and image quality evaluation methods are used to design the macro-optical structure parameters of the Offner system. And then the incident ray information, such as the incident angle and the polarization states are calculated by using the three-dimensional polarization ray-tracing method. To improve the diffraction efficiency, we combine rigorous coupled wave analysis and a particle swarm optimization algorithm to optimize the microstructure parameters of the convex-blazed grating. Further, a convex-blazed grating in a mid-wave infrared Offner imaging spectrometer is designed as an example to illustrate our design method in detail. The design results indicate that the Offner imaging spectrometer has good imaging quality, and the average diffraction efficiency of the -1st diffraction order of the convex-blazed grating in the spectral coverage 3-5 µm is 82.24%. Compared to the traditional design method, the lowest spectral diffraction efficiency is improved from 59.88% to 69.24%, the highest spectral diffraction efficiency is improved from 90.45% to 91.84%, and the standard deviation is reduced from 7.82 to 6.62.

Comparison of CT Image Performance with or without Tin Filter based on Blind Image Quality Evaluation Method

Comparison of CT Image Performance with or without Tin Filter based on Blind Image Quality Evaluation Method

Noise reduction for sonar images by statistical analysis and fields of experts

Sonar images are usually suffering from speckle noise which results in poor visual quality. In order to improve the sonar imaging quality, removing or reducing these speckle noises is a very important and arduous task. In this paper, the imaging principle and noise characteristics of the side-scan sonar (SSS) are analyzed, and five typical probability distribution functions are used to fit the seabed reverberation. Through experiment comparison, the Gamma distribution is selected to simulate the noise of the SSS image caused by the reverberation. Simultaneously, the fields of experts denoising algorithm based on the Gamma distribution (Gamma FoE) is proposed for SSS image denoising. In order to perceive and measure the denoising effect better, evaluation indexes of Fast Noise Variance Estimation (FNVE, an image noise estimation method) and Blind Referenceless Image Spatial Quality Evaluator (BRISQUE, an image quality evaluation method) are selected for image quality perception. The final results of the SSS image denoise experiment show that the Gamma FoE denoise algorithm has a better effect on SSS image denoise application than other denoise algorithms.

Analysis and Optimization of Image Quality Evaluation Function in Computer-aided Alignment of Precision Optical System

The consistent application of computer-aided alignment technology could surmount aimlessness and significantly shorten the align period of complex optical systems. Wherein, the appropriate system image evaluation function is the primary condition to solve the system misalignment successfully. This paper first establishes the relationship between the amount of misalignment and the system aberration through mathematical modelling and then analyzes the relationship between Zernike coefficient, geometric aberration and system image quality in detail, proposes to construct a system image evaluation system using the idea of Zernike coefficient weighting to solve the problem of numerical instability and imprecision of the solutions caused by the direct selection of Zernike coefficients. Subsequently, the experimental alignment of an infrared optical system was performed using the optical system image quality evaluation method established in this paper. The alignment results showed that the RMS value of the system image quality improved by 24.8% under the same solving algorithm compares to the direct selection of the Zernike coefficient and the weighted Zernike coefficient as the system image quality evaluation function. Therefore, the application of the weighted Zernike coefficient-based optical system image evaluation function proposed in this paper is reasonable, feasible, and effective in the computer-aided alignment process.

Research on Image Quality Evaluation Method of Depth Learning Model Based on Core Concept

In view of the fact that the existing image quality evaluation methods are generally artificial design features, it is difficult to automatically and effectively extract image features that conform to the human visual system. Inspired by human visual characteristics, a new full reference image quality evaluation method based on depth learning model based on core concepts is proposed. Firstly, depth learning algorithm is used to extract multi-layer features from reference images and distorted images respectively. Then, the local similarity of the feature map of the reference image and the distorted image in each layer is calculated as the local quality description of the corresponding depth. Finally, the local quality of all layers is synthesized to obtain the overall quality score of the image. On the basis of the pre-training model, the depth model network is fine-tuned by using the image visual evaluation data set to obtain a depth model for evaluation. The standard experiment shows that fine-tuning training of each pre-training model on the standard data set achieves good classification results. Experiments show that the designed depth learning model based on core concepts is superior to the existing full reference image quality evaluation methods, and its prediction results have good accuracy and consistency with subjective quality evaluation.

Evaluating medical images using deep convolutional neural networks: A simulated CT phantom image study.

Applied research on artificial intelligence, mainly in deep learning, is widely performed. If medical images can be evaluated using artificial intelligence, this could substantially improve examination efficiency. We investigated an evaluation system for medical images with different noise characteristics using a deep convolutional neural network. Simulated computed tomography images are the targets of the system. We used an AlexNet trained with natural images for the deep convolutional neural network and a support vector machine for classification. Synthetic computed tomography images with circular and rectangular signal bodies at different levels of contrast and added Gaussian noise were used for training and testing. Two transfer learning methods were tested: classification by a re-trained support vector machine using the AlexNet features, and a method that fine-tuned the deep convolutional neural network. Using the first method, all the test image noise levels could be classified correctly. The fine-tuning method achieved an accuracy rate of 92.6%. An image quality evaluation method using artificial intelligence will be useful for clinical images and different image quality indices in the future.

Hyperspectral image quality evaluation using generalized regression neural network

In order to alleviate the overfitting problem caused by image quality evaluation (IQA) model learning under intolerably small dataset, this paper proposes a multi-feature fusion-based deep architecture for hyperspectral image quality assessment. First, eight key IQA-related features, which are descriptive to the mean noise of multi-band images, spatial correlation, inter-spectral correlation, blur, and the phase-consistent map of images, are extracted from each hyperspectral image within the dataset. Based on this, a carefully-designed generalized regression neural network (GRNN) with a limited number of parameters is hierarchically trained by the feature vectors from samples in the training IQA data set. Comprehensive experimental evaluations on the hyperspectral IQA images from the DOTA dataset and the EO-1 Hyperion dataset have shown that the proposed model can indicate the subjective/objective quality-aware images regions precisely In addition, we observe that our designed IQA method has received impressive IQA performance than the other state-of-the-art non-reference methods.

Study of the depth accuracy and entropy characteristics of a ToF camera with coupled noise

By illuminating targets with a modulated light signal, a time-of-flight (ToF) camera can calculate depth maps based on the phase shifts generated by the round-trip travel of the light from the targets back to the sensor. The accuracy of such depth imaging depends on the quality of the returned light. However, the calculation can be disturbed by various factors, including the external environment and the internal structure of the camera. Multiple coupled interferences can introduce noise into the depth data collected by a ToF camera in the spatial domain. It is difficult to express the relationship between the noise in depth maps and these mixed disturbances in a mathematical model. Based on the theory of differential entropy and a large amount of depth data from a ToF camera, this paper analyzes the characteristics of depth imaging entropy, proposes an evaluation method for depth image quality, and presents a multilayer perceptron model with information entropy (E-MLP) trained to optimize the accuracy of depth imaging. Experimental results show that this method can significantly improve the depth accuracy in the case of mixed noise.

Optical Flow Estimation and Denoising of Video Images Based on Deep Learning Models

In order to effectively extract image features highly related to visual perception quality, improve the image quality evaluation method, under the framework of deep learning, combining the optical flow method and the edge detection algorithm, a multi-feature fusion motion based on improved optical flow is proposed target detection algorithm. First, a video fluid model is proposed. The fluid model decomposes the video object area changes into sub-area zoom, rotation and translation movements. The rigid body area and the area hierarchy describe the spatial relationship of pixels, and the rigid body motion describes the time domain relationship of pixels. It provides a region-based video processing. The associated spatiotemporal is association method. Secondly, a video fluid model is proposed. The video fluid model treats all pixels of the same surface imaged in the video as a fluid, using streamlines to represent the regional motion of the object, and streamlines to represent the pixel motion of the video object, using rotators and translation lines to simplify the streamlines when necessary. The streamline of the same fluid is smooth in the time domain, and the flow pattern is smooth in both the time domain and the space domain. Finally, the top-down deep learning generation model conversion is carried out, and finally through continuous adjustment between different levels, the generation model can reconstruct the original sample with lower error, so that the essential characteristics of this sample are obtained, namely the highest abstract representation of the depth model. After processing the deep learning model, the sample features after dimensionality reduction can be obtained, and the recognition module is used on this basis. Experiments show that the optical flow estimation method based on deep learning and multi-grid, optical flow field estimation method based on variational model and desiccation method proposed in this paper are effective, and it is suitable for moving image analysis, target tracking and 3D reconstruction Such research has certain theoretical significance and practical application value.

Difference of Visual Information Metric Based on Entropy of Primitive

Image sparse representation is a method of efficient compression and coding of image signal in the process of digital image processing. Image after sparse representation, to enhance the transmission efficiency of the image signal. Entropy of Primitive (EoP) is a statistical representation of the sparse representation of the image, which indicates the probability of each base element. Based on the EoP, this paper presents an image quality evaluation method-Difference of Visual Information Metric (DVIM). The principle of this method is to evaluate the image quality with the difference between the original image and the distorted image. The comparative experiments between DVIM & PSNR & SSIM are carried out. It was found that there was a great improvement in the image quality evaluation of geometric changes. This method is an effective image quality evaluation method, which overcomes the weakness of other quality evaluation methods for geometrically changing images to a certain extent, and is more consistent with the subjective observation of the human eye.

Nonreference Image Quality Evaluation Algorithm Based on Wavelet Convolutional Neural Network and Information Entropy

The image quality evaluation method, based on the convolutional neural network (CNN), achieved good evaluation performance. However, this method can easily lead the visual quality of image sub-blocks to change with the spatial position after the image is processed by various distortions. Consequently, the visual quality of the entire image is difficult to reflect objectively. On this basis, this study combines wavelet transform and CNN method to propose an image quality evaluation method based on wavelet CNN. The low-frequency, horizontal, vertical, and diagonal sub-band images decomposed by wavelet transform are selected as the inputs of convolution neural network. The feature information in multiple directions is extracted by convolution neural network. Then, the information entropy of each sub-band image is calculated and used as the weight of each sub-band image quality. Finally, the quality evaluation values of four sub-band images are weighted and fused to obtain the visual quality values of the entire image. Experimental results show that the proposed method gains advantage from the global and local information of the image, thereby further improving its effectiveness and generalization.

Research on image quality in decision management system and information system framework

With the continuous development of society and the rapid advancement of science and technology, all walks of life are increasingly required to effectively manage their projects or affairs. Decision-making is the basis of project management, and the formation of decisions determines the formulation of corresponding programs, organizations and businesses. Decision-making is generated by decision makers based on the multi-faceted information they have, so decision management systems and information systems are inseparable in the management of projects and transactions. At present, research on decision management systems emphasizes management and decisions, and few researchers combine decision management and related information. In view of the current research deficiencies, and taking into account the needs of various industries for image applications, this paper introduces image quality evaluation methods, and systematically studies the decision management system and information system framework. The experiment part takes the intelligent image recognition and prevention system of citrus pests and diseases as an example, and the role of image quality in decision management system and information system from the perspective of practice. The results show that through the image quality evaluation system, in the identification and prevention of citrus pests and diseases, it can combine various aspects of information to make effective prevention and control decisions. The results show that the introduction of image quality evaluation can improve the comprehensiveness of information system. On this basis, decision-making system can help decision makers make correct decisions quickly and effectively.