Increase Image Resolution Research Articles

Using super-resolution generative adversarial network models and transfer learning to obtain high resolution digital periapical radiographs

Periapical Radiographs are commonly used to detect several anomalies, like caries, periodontal, and periapical diseases. Even considering that digital imaging systems used nowadays tend to provide high-quality images, external factors, or even system limitations can result in a vast amount of radiographic images with low quality and resolution. Commercial solutions offer tools based on interpolation methods to increase image resolution. However, previous literature shows that these methods may create undesirable effects in the images affecting the diagnosis accuracy. One alternative is using deep learning-based super-resolution methods to achieve better high-resolution images. Nevertheless, the amount of data for training such models is limited, demanding transfer learning approaches. In this work, we propose the use of super-resolution generative adversarial network (SRGAN) models and transfer learning to achieve periapical images with higher quality and resolution. Moreover, we evaluate the influence of using the transfer learning approach and the datasets selected for it in the final generated images. For that, we performed an experiment comparing the performance of the SRGAN models (with and without transfer learning) with other super-resolution methods. Considering Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Mean Opinion Score (MOS), the results of SRGAN models using transfer learning were better on average. This superiority was also verified statistically using the Wilcoxon paired test. In the visual analysis, the high quality achieved by the SRGAN models, in general, is visible, resulting in more defined edges details and fewer blur effects.

Frontiers of Sodium MRI Revisited: From Cartilage to Brain Imaging.

Sodium magnetic resonance imaging (23Na‐MRI) is a highly promising imaging modality that offers the possibility to noninvasively quantify sodium content in the tissue, one of the most relevant parameters for biochemical investigations. Despite its great potential, due to the intrinsically low signal‐to‐noise ratio (SNR) of sodium imaging generated by low in vivo sodium concentrations, low gyromagnetic ratio, and substantially shorter relaxation times than for proton (1H) imaging, 23Na‐MRI is extremely challenging. In this article, we aim to provide a comprehensive overview of the literature that has been published in the last 10–15 years and which has demonstrated different technical designs for a range of 23Na‐MRI methods applicable for disease diagnoses and treatment efficacy evaluations. Currently, a wider use of 3.0T and 7.0T systems provide imaging with the expected increase in SNR and, consequently, an increased image resolution and a reduced scanning time. A great interest in translational research has enlarged the field of sodium MRI applications to almost all parts of the body: articular cartilage tendons, spine, heart, breast, muscle, kidney, and brain, etc., and several pathological conditions, such as tumors, neurological and degenerative diseases, and others. The quantitative parameter, tissue sodium concentration, which reflects changes in intracellular sodium concentration, extracellular sodium concentration, and intra–/extracellular volume fractions is becoming acknowledged as a reliable biomarker. Although the great potential of this technique is evident, there must be steady technical development for 23Na‐MRI to become a standard imaging tool. The future role of sodium imaging is not to be considered as an alternative to 1H MRI, but to provide early, diagnostically valuable information about altered metabolism or tissue function associated with disease genesis and progression.Level of Evidence1Technical Efficacy Stage1

SUPER-RESOLUTION OF MULTISPECTRAL SATELLITE IMAGES USING CONVOLUTIONAL NEURAL NETWORKS

Abstract. Super-resolution aims at increasing image resolution by algorithmic means and has progressed over the recent years due to advances in the fields of computer vision and deep learning. Convolutional Neural Networks based on a variety of architectures have been applied to the problem, e.g. autoencoders and residual networks. While most research focuses on the processing of photographs consisting only of RGB color channels, little work can be found concentrating on multi-band, analytic satellite imagery. Satellite images often include a panchromatic band, which has higher spatial resolution but lower spectral resolution than the other bands. In the field of remote sensing, there is a long tradition of applying pan-sharpening to satellite images, i.e. bringing the multispectral bands to the higher spatial resolution by merging them with the panchromatic band. To our knowledge there are so far no approaches to super-resolution which take advantage of the panchromatic band. In this paper we propose a method to train state-of-the-art CNNs using pairs of lower-resolution multispectral and high-resolution pan-sharpened image tiles in order to create super-resolved analytic images. The derived quality metrics show that the method improves information content of the processed images. We compare the results created by four CNN architectures, with RedNet30 performing best.

Image enhancement with DAGNet-based artificial neural network

The article is devoted to an artificial neural network that increases image resolution. According to the meaning of the article, two parts can be distinguished. The first part of the article discusses the process of designing an artificial neural network using DAGNet technology, during which its main parameters and structure are described. The second part is devoted to assessing the quality of functioning of the developed artificial neural network. To evaluate the performance of ANNs a comparative analysis with the bicubic interpolation method was used.

Organic Spherical Nucleic Acids for the Transport of a NIR‐II‐Emitting Dye Across the Blood–Brain Barrier

AbstractDNA nanotechnology plays an increasingly important role in the biomedical field; however, its application in the design of organic nanomaterials is underexplored. Herein, we report the use of DNA nanotechnology to transport a NIR‐II‐emitting nanofluorophore across the blood–brain barrier (BBB), facilitating non‐invasive imaging of brain tumors. Specifically, the DNA block copolymer, PS‐b‐DNA, is synthesized through a solid‐phase click reaction. We demonstrate that its self‐assembled structure shows exceptional cluster effects, among which BBB‐crossing is the most notable. Therefore, PS‐b‐DNA is utilized as an amphiphilic matrix to fabricate a NIR‐II nanofluorephore, which is applied in in vivo bioimaging. Accordingly, the NIR‐II fluorescence signal of the DNA‐based nanofluorophore localized at a glioblastoma is 3.8‐fold higher than the NIR‐II fluorescence signal of the PEG‐based counterpart. The notably increased imaging resolution will significantly benefit the further diagnosis and therapy of brain tumors.

Organic Spherical Nucleic Acids for the Transport of a NIR-II-Emitting Dye Across the Blood-Brain Barrier.

DNA nanotechnology plays an increasingly important role in the biomedical field; however, its application in the design of organic nanomaterials is underexplored. Herein, we report the use of DNA nanotechnology to transport a NIR-II-emitting nanofluorophore across the blood-brain barrier (BBB), facilitating non-invasive imaging of brain tumors. Specifically, the DNA block copolymer, PS-b-DNA, is synthesized through a solid-phase click reaction. We demonstrate that its self-assembled structure shows exceptional cluster effects, among which BBB-crossing is the most notable. Therefore, PS-b-DNA is utilized as an amphiphilic matrix to fabricate a NIR-II nanofluorephore, which is applied in in vivo bioimaging. Accordingly, the NIR-II fluorescence signal of the DNA-based nanofluorophore localized at a glioblastoma is 3.8-fold higher than the NIR-II fluorescence signal of the PEG-based counterpart. The notably increased imaging resolution will significantly benefit the further diagnosis and therapy of brain tumors.

Applications of dielectric pads, novel materials and resonators in 1.5T and 3T MRI

In order to boost the performance of magnetic resonance imaging without increasing the static magnetic field, it is necessary to increase its intrinsic sensitivity. This allows a reduction in the scanning time, increased spatial resolution, and can enable low-field strength systems (which are much cheaper and can be used to scan patients with metallic implants) to have a higher signal-to-noise ratio (SNR) so that they are comparable to more expensive higher field strength systems. In this contribution, we demonstrate radiofrequency field enhancing and shaping devices based on novel materials, such as high permittivity dielectric structures and metamaterials. These materials can substantially enhance SNR, thus potentially increasing image resolution or allowing faster examinations.

Optimizing Infrared Camera Resolution for Small Object Detection using Subpixel Rendering and PIFS in Multiresolution Image Analysis.

Background: Breast cancer screening techniques have been developing rapidly in the field of imaging systems. One of these techniques is thermography, which is an alternative modality for mammography to detect breast lesions. Thermography utilization has been progressively developing as various models and methods of object processing improvement. Currently, the Fluke TIS20 infrared camera, with a resolution of 320 × 240, has not been used to measure precisely small objects such as early breast cancer lesions. Retrieval and processing of single images lead into imprecise object measurements and false positive results.Objective: Problems have been arisen due to the limitations of the camera resolution, object retrieval techniques and suboptimal image processing. The aim of this study was to detect accurately breast cancer lesions in rats, which were induced by carcinogenic compounds.Material and Methods: In this experimental study, development of models was conducted based on increasing image by optimizing the ability of low-resolution infrared (IR) cameras to identify s mall objects precisely. Image pixel density increased by adjusting the distance of the objects from the camera and multiple images of objects gradually shifting were used to measure object dimensions precisely.Results: The results showed that cancerous lesions as small as 1.27 mm could be detected. This method of lesion detection had a sensitivity and specificity of 93% and 77 % respectively.Conclusion: Small objects (cancerous lesions) were measured by increasing image resolution through splitting pixels into subpixels and combining several images using Partitioned Iterated Function Systems (PIFS).

Resizing and cleaning of histopathological images using generative adversarial networks

Bilinear and Bicubic interpolation techniques are frequently used to increase image resolution. These techniques with data modeling approach are replaced by intelligent systems that can learn automatically from data. SRGAN is a modern Generative Adversarial Network developed as an alternative to classical interpolation techniques. His ability to produce images in super resolution has attracted the attention of many researchers.In this study, noise elimination performance of super resolution generative adversarial network (SRGAN) with image magnification was investigated. The results of the noise cleaning were compared with the classical approaches (mean, median, adaptive filters). SSIM, PSNR and FFT_MSE metrics were evaluated in experimental studies using images in the data set Camelyon17. When the results were evaluated, it was observed that SRGAN was superior to the classical approaches not only in increasing the resolution but also in the noise cleaning area.

Measurements of cerebral blood volume using quantitative susceptibility mapping, R2 * relaxometry, and ferumoxytol-enhanced MRI.

Ferumoxytol-enhanced MRI holds potential for the non-invasive assessment of vascular architecture using estimates of cerebral blood volume (CBV). Ferumoxytol specifically enables steady-state imaging with extended acquisition times, for substantial improvements in resolution and contrast-to-noise ratio. With such data, quantitative susceptibility mapping (QSM) can be used to obtain images of local tissue magnetic susceptibility and hence estimate the increase in blood susceptibility after administration of a contrast agent, which in turn can be correlated to tissue CBV. Here, we explore the use of QSM for CBV estimation and compare it with R2 * (1/T2 *)-based results. Institutional review board approval was obtained, and all subjects provided written informed consent. For this prospective study, MR images were acquired on a 3.0T scanner in 19 healthy subjects using a multiple-echo T2 *-weighted sequence. Scanning was performed before and after the administration of two doses of ferumoxytol (1mg FE/kg and 4mg FE/kg). Different QSM approaches were tested on numerical phantom simulations. Results showed that the accuracy of magnetic susceptibility measurements improved with increasing image resolution and decreasing vascular density. In vivo changes in magnetic susceptibility were measured after the administration of ferumoxytol utilizing QSM, and significantly higher QSM-based CBV was measured in gray matter compared with white matter. QSM- and R2 *-based CBV estimates correlated well, with similar average values, but a larger variance was found in QSM-based estimates.

Static volumetric three-dimensional display based on an electric-field-controlled two-dimensional optical beam scanner.

Using the quadratic electro-optic effect and the gradient of the composition ratio m [Nb/(Ta+Nb) in mol. %] in a potassium tantalate niobate crystal, we have designed an electric-field-controlled two-dimensional optical beam scanner with a wide wavelength range and fast response. This scanner is used to realize a volumetric display based on a two-frequency, two-step upconversion technique that is used to address the imaging volume. Use of appropriately designed imaging optics and custom-designed software to convert 2D renderings of volumetric images into control signals for the scanner along with appropriate infrared laser source selection allows efficient single-color image generation with a large viewing zone, without flicker and with natural depth cues. The resulting system has the potential to increase image resolution to nearly 61.5×109 with high scanning frequency and to expand to display three-color imagery.

Study of Hot Spots Probing Based on Meta-Pillars Array

Inspired by imaging principle of near-field scanning optical microscope (NSOM), meta-pillars array is designed and analyzed on the basis of microscopic imaging application with high resolution. Finely focused spots acting as tiny secondary sources for illumination at near-field can be derived under supporting of the meta-pillars for the purpose of increasing imaging resolution. Numerical calculation is carried out on the basis of finite difference and time domain (FDTD) algorithm. Our calculation results demonstrate that the meta-pillars are capable of supporting the microscopic imaging at sub-wavelength resolution.

Handheld multi-frame super-resolution

Compared to DSLR cameras, smartphone cameras have smaller sensors, which limits their spatial resolution; smaller apertures, which limits their light gathering ability; and smaller pixels, which reduces their signal-to-noise ratio. The use of color filter arrays (CFAs) requires demosaicing, which further degrades resolution. In this paper, we supplant the use of traditional demosaicing in single-frame and burst photography pipelines with a multiframe super-resolution algorithm that creates a complete RGB image directly from a burst of CFA raw images. We harness natural hand tremor, typical in handheld photography, to acquire a burst of raw frames with small offsets. These frames are then aligned and merged to form a single image with red, green, and blue values at every pixel site. This approach, which includes no explicit demosaicing step, serves to both increase image resolution and boost signal to noise ratio. Our algorithm is robust to challenging scene conditions: local motion, occlusion, or scene changes. It runs at 100 milliseconds per 12-megapixel RAW input burst frame on mass-produced mobile phones. Specifically, the algorithm is the basis of the Super-Res Zoom feature, as well as the default merge method in Night Sight mode (whether zooming or not) on Google's flagship phone.

Synthetic-aperture spatial light modulation with interferometric correction of the point spread function

Holographic projection based on a collective matrix of spatial light modulators (synthetic aperture SLMs) allows scalable increase of image resolution for the price of strong side-lobes in the point spread function spots. This may cause problems, for example, in optical trapping where spurious straps are potentially formed. Here we present a simple method of removing such side-lobes with the use of third coherent beam which destructively interferes with the unwanted fields at the image plane. Theoretical background, numerical simulations and experimental validation are presented.

A novel method for extracting information on pores from cast thin-section images

In rock physics and petrological applications, pore identification from cast thin-section (CTS) images is a widely used means of estimating porosity and evaluating types of pores and pore structure parameters. Common problems encountered in current automatic methods of pore extraction from these images are accuracy and/or computational efficiency, especially as the image resolution increases. In high resolution, the transition boundaries between pores and matrices can easily be wrongly identified by automatic extraction methods, which would have significant impacts on the final pore estimation. To address this problem, we propose a revised multiple threshold method combined with error correction and image refining. The method, named ctsPore, is implemented in the hue-saturation-value colour space and comprises three steps: coarse extraction of pores from thin-section images; removal or reduction of incorrectly extracted pores on the surface of identified particle grains and within the transition boundaries between pores and rock matrices; and refinement of extracted pore images by removing unrealistically small areas within identified particle grains or pore regions. The first step applies the threshold method based on the hues of the pixels; the second step is based on the product of saturation and value; and the third step is based on the statistics of small areas. To demonstrate the application of the proposed method, a series of comparison studies were conducted using cyan-, blue- and magenta-impregnated cast thin-section images from the southern margin of the Junggar Basin, China. The results show that ctsPore is an accurate and efficient means of extracting the pore information from high-resolution CTS images impregnated with different colour agents.

Emerging Magnetic Resonance Imaging Techniques for Atherosclerosis Imaging.

Atherosclerosis is a prevalent disease affecting a large portion of the population at one point in their lives. There is an unmet need for noninvasive diagnostics to identify and characterize at-risk plaque phenotypes noninvasively and in vivo, to improve the stratification of patients with cardiovascular disease, and for treatment evaluation. Magnetic resonance imaging is uniquely positioned to address these diagnostic needs. However, currently available magnetic resonance imaging methods for vessel wall imaging lack sufficient discriminative and predictive power to guide the individual patient needs. To address this challenge, physicists are pushing the boundaries of magnetic resonance atherosclerosis imaging to increase image resolution, provide improved quantitative evaluation of plaque constituents, and obtain readouts of disease activity such as inflammation. Here, we review some of these important developments, with specific focus on emerging applications using high-field magnetic resonance imaging, the use of quantitative relaxation parameter mapping for improved plaque characterization, and novel 19F magnetic resonance imaging technology to image plaque inflammation.

A watermarking algorithm for image content authentication in double-compression environment

Malicious tampering of digital image content may lead to serious consequences. As image resolution increases and its circulations on the network grows rapidly, images undergo double compression in the following environment: compression (JPEG/JPEG2000) for first release, decoding, application processing (conventional signal processing/security attacks/malicious content tampering), and compression (JPEG2000/JPEG) for rerelease before detection. Effectively determining whether the image content has been tampered and to locate the tampering position in a double-compression environment is an urgent problem. In this paper, a novel watermark expression method based on rotating vector and its modulation algorithm is proposed. Based on this method, a semi-fragile watermarking scheme is established for the image content authentication. The stability of watermarked data is analyzed theoretically, providing the watermark scheme with a theoretical basis. The authentication schemes with feature extraction and reconstruction, watermark embedding and extraction, and tamper detection and localization are mainly elaborated. The robustness of watermark, its security, and related detection performance are analyzed. Theoretical analysis and experiments show that the scheme features a good watermark transparency, good robustness, and stable distribution for different attacks and can effectively distinguish malicious tampering from content-preserving operators and locate tampering area under double-compression environment. In addition, the proposed scheme can resist watermark attacks, collage attacks, and forge attacks. Compared with related schemes, the proposed scheme has superior comprehensive performance and is suitable for content authentication in double-compression environment, which expands the application scope of watermark-based content authentication.

A Resolution-Free Parallel Algorithm for Image Edge Detection within the Framework of Enzymatic Numerical P Systems.

Image edge detection is a fundamental problem in image processing and computer vision, particularly in the area of feature extraction. However, the time complexity increases squarely with the increase of image resolution in conventional serial computing mode. This results in being unbearably time consuming when dealing with a large amount of image data. In this paper, a novel resolution free parallel implementation algorithm for gradient based edge detection, namely EDENP, is proposed. The key point of our method is the introduction of an enzymatic numerical P system (ENPS) to design the parallel computing algorithm for image processing for the first time. The proposed algorithm is based on a cell-like P system with a nested membrane structure containing four membranes. The start and stop of the system is controlled by the variables in the skin membrane. The calculation of edge detection is performed in the inner three membranes in a parallel way. The performance and efficiency of this algorithm are evaluated on the CUDA platform. The main advantage of EDENP is that the time complexity of can be achieved regardless of image resolution theoretically.

Boosting magnetic resonance imaging signal-to-noise ratio using magnetic metamaterials

Magnetic resonance imaging (MRI) represents a mainstay among the diagnostic imaging tools in modern healthcare. Signal-to-noise ratio (SNR) represents a fundamental performance metric of MRI, the improvement of which may be translated into increased image resolution or decreased scan time. Recently, efforts towards the application of metamaterials in MRI have reported improvements in SNR through their capacity to interact with electromagnetic radiation. While promising, the reported applications of metamaterials to MRI remain impractical and fail to realize the full potential of these unique materials. Here, we report the development of a magnetic metamaterial enabling a marked boost in radio frequency field strength, ultimately yielding a dramatic increase in the SNR (~ 4.2×) of MRI. The application of the reported magnetic metamaterials in MRI has the potential for rapid clinical translation, offering marked enhancements in SNR, image resolution, and scan efficiency, thereby leading to an evolution of this diagnostic tool.

An object-based and heterogeneous segment filter convolutional neural network for high-resolution remote sensing image classification

ABSTRACTIn recent years, object-based segmentation methods and shallow-model classification algorithms have been widely integrated for remote sensing image supervised classification. However, as the image resolution increases, remote sensing images contain increasingly complex characteristics, leading to higher intraclass heterogeneity and interclass homogeneity and thus posing substantial challenges for the application of segmentation methods and shallow-model classification algorithms. As important methods of deep learning technology, convolutional neural networks (CNNs) can hierarchically extract higher-level spatial features from images, providing CNNs with a more powerful recognition ability for target detection and scene classification in high-resolution remote sensing images. However, the input of the traditional CNN is an image patch, the shape of which is scarcely consistent with a given segment. This inconsistency may lead to errors when directly using CNNs in object-based remote sensing classification: jagged errors may appear along the land cover boundaries, and some land cover areas may overexpand or shrink, leading to many obvious classification errors in the resulting image. To address the above problem, this paper proposes an object-based and heterogeneous segment filter convolutional neural network (OHSF-CNN) for high-resolution remote sensing image classification. Before the CNN processes an image patch, the OHSF-CNN includes a heterogeneous segment filter (HSF) to process the input image. For the segments in the image patch that are obviously different from the segment to be classified, the HSF can differentiate them and reduce their negative influence on the CNN training and decision-making processes. Experimental results show that the OHSF-CNN not only can take full advantage of the recognition capabilities of deep learning methods but also can effectively avoid the jagged errors along land cover boundaries and the expansion/shrinkage of land cover areas originating from traditional CNN structures. Moreover, compared with the traditional methods, the proposed OHSF-CNN can achieve higher classification accuracy. Furthermore, the OHSF-CNN algorithm can serve as a bridge between deep learning technology and object-based segmentation algorithms thereby enabling the application of object-based segmentation methods to more complex high-resolution remote sensing images.