Super-resolution Concept Research Articles

Improving Resolution of Panoramic Radiographs: Super Resolution Concept.

Improving Resolution of Panoramic Radiographs: Super Resolution Concept.

Super-Resolution DOA Estimation for Wideband Signals Using Non-Uniform Linear Arrays With No Focusing Matrix

We focus on developing an effective Direction Of Arrival (DOA) estimation method for wideband sources based on the super-resolution concept. In this letter, we represent the data model of a wideband DOA estimation problem in a new form. Then, using this model, we propose an atomic norm minimization problem that leads to an accurate wideband DOA estimation method with no need for any focusing matrices. Moreover, this method requires no initial estimate and can be implemented by non-uniform linear arrays. Numerical simulations show that in comparison to some well-known techniques, the proposed method generates outstanding accuracy and higher robustness to noise. The effectiveness of the method is also verified in presence of close adjacent sources.

LSRN: A Recurrent Residual Learning Framework for Continuous Wireless Channel Estimation Using Super-Resolution Concept

As only a few parts of wireless resources can be utilized for pilot transmission, channel estimation, especially the interpolation process, has often been recognized as a challenging ill-posed reconstruction problem. To deal with this task, we formulate it as a typical image super resolution problem, and propose a recurrent residual learning framework named LSRN. Our proposed scheme jointly utilizes the advantages of recurrent and residual structure in the machine learning area to approximate the non-linear interpolation relations between the reference signal and surrounding resource elements. In addition, we propose a low complexity implementation scheme called LSRN-L to address the stringent processing delay requirement in the channel estimation tasks. Through numerical examples as well as prototype verification, the proposed LSRN/LSRN-L can easily outperform the convolutional GI plus DFT based interpolation scheme by 10dB in terms of normalized mean square error. Meanwhile, the low complexity LSRN-L can maintain the processing delay within one millisecond.

Super-Resolution for Monocular Depth Estimation With Multi-Scale Sub-Pixel Convolutions and a Smoothness Constraint

Depth estimation from a monocular image is of paramount importance in various vision tasks, such as obstacle detection, robot navigation, and 3D reconstruction. However, how to get an accurate depth map with clear details and a fine resolution remains an unresolved issue. As an attempt to solve this problem, we exploit image super-resolution concepts and techniques for monocular depth estimation and propose a novel CNN-based approach, namely $MSCN_{NS}$ , which involves multi-scale sub-pixel convolutions and a neighborhood smoothness constraint. Specifically, $MSCN_{NS}$ makes use of sub-pixel convolutions with multi-scale fusions to retrieve a high-resolution depth map with fine details of the scene. Different from previous multi-scale fusion strategies, those multi-scale features come from supervised scale branches of the network. Furthermore, $MSCN_{NS}$ incorporates a neighborhood smoothness regularization term to make sure that spatially closer pixels with similar features would have close depth values. The effectiveness and efficiency of $MSCN_{NS}$ have been corroborated through extensive experiments conducted on benchmark datasets.

Leveraging multi-layer imager detector design to improve low-dose performance for megavoltage cone-beam computed tomography

While megavoltage cone-beam computed tomography (CBCT) using an electronic portal imaging device (EPID) provides many advantages over kilovoltage (kV) CBCT, clinical adoption is limited by its high doses. Multi-layer imager (MLI) EPIDs increase DQE(0) while maintaining high resolution. However, even well-designed, high-performance MLIs suffer from increased electronic noise from each readout, degrading low-dose image quality. To improve low-dose performance, shift-and-bin addition (ShiBA) imaging is proposed, leveraging the unique architecture of the MLI. ShiBA combines hardware readout-binning and super-resolution concepts, reducing electronic noise while maintaining native image sampling. The imaging performance of full-resolution (FR); standard, aligned binned (BIN); and ShiBA images in terms of noise power spectrum (NPS), electronic NPS, modulation transfer function (MTF), and the ideal observer signal-to-noise ratio (SNR)—the detectability index (d′)—are compared. The FR 4-layer readout of the prototype MLI exhibits an electronic NPS magnitude 6-times higher than a state-of-the-art single layer (SLI) EPID. Although the MLI is built on the same readout platform as the SLI, with each layer exhibiting equivalent electronic noise, the multi-stage readout of the MLI results in electronic noise 50% higher than simple summation. Electronic noise is mitigated in both BIN and ShiBA imaging, reducing its total by ~12 times. ShiBA further reduces the NPS, effectively upsampling the image, resulting in a multiplication by a sinc2 function. Normalized NPS show that neither ShiBA nor BIN otherwise affects image noise. The LSF shows that ShiBA removes the pixilation artifact of BIN images and mitigates the effect of detector shift, but does not quantifiably improve the MTF. ShiBA provides a pre-sampled representation of the images, mitigating phase dependence. Hardware binning strategies lower the quantum noise floor, with 2 × 2 implementation reducing the dose at which DQE(0) degrades by 10% from 0.01 MU to 0.004 MU, representing 20% improvement in d′.

Pushing back the limits of Raman imaging by coupling super-resolution and chemometrics for aerosols characterization.

The increasing interest in nanoscience in many research fields like physics, chemistry, and biology, including the environmental fate of the produced nano-objects, requires instrumental improvements to address the sub-micrometric analysis challenges. The originality of our approach is to use both the super-resolution concept and multivariate curve resolution (MCR-ALS) algorithm in confocal Raman imaging to surmount its instrumental limits and to characterize chemical components of atmospheric aerosols at the level of the individual particles. We demonstrate the possibility to go beyond the diffraction limit with this algorithmic approach. Indeed, the spatial resolution is improved by 65% to achieve 200 nm for the considered far-field spectrophotometer. A multivariate curve resolution method is then coupled with super-resolution in order to explore the heterogeneous structure of submicron particles for describing physical and chemical processes that may occur in the atmosphere. The proposed methodology provides new tools for sub-micron characterization of heterogeneous samples using far-field (i.e. conventional) Raman imaging spectrometer.

Digital holographic microscope for measurement of high gradient deep topography object based on superresolution concept

In this Letter, a novel concept based on superresolution technique that enables the measurement of high gradient and deep topography objects using digital holographic (DH) microscopy is introduced. The major problem of DH systems is limited NA that prohibits the metrological characterization of object features of high frequencies. The proposed technique has the ability to extend spatial frequency spectrum of the measured topography by applying multidirectional plane wave illumination, which is experimentally realized with a grating. The technique recovers sample topography from the set of object waves with different object spectra that are converted into a set of topographies by using an algorithm which takes into account refraction. Application of this novel approach is experimentally validated by characterization of high gradient topography objects with maximum angle of tangent 65°.

Super-Resolved Raman Spectroscopy

ABSTRACT In this paper we present a super-resolved Raman spectroscopy configuration where, due to the proposed resolution improvement, more chemical species can be simultaneously detected. The idea includes usage of the basic property of Raman in which shifting the excitation wavelength will shift the absolute wavelength of the Raman spectral lines. Thus, encoding the spectra prior to analysis by the spectrometer allows obtaining spectral super-resolved sensing via time multiplexing super-resolution concepts applied over the spectral domain.

On the resolution of lenses made of a negative-index material

Resolution of the lenses made of a negative-index material is considered. It is shown that the super-resolution concept is untenable and the possibility of obtaining a perfect image on its own eventually contradicts Maxwell's equations in vacuum. It is also shown that known limitations of the diffraction theory on resolution of optical instruments hold true for the resolution of lenses of a negative-index material, in particular, the resolution of a Veselago lens.

Increasing the spatial resolution of near infrared chemical images (NIR-CI): The super-resolution paradigm applied to pharmaceutical products

Near-infrared chemical imaging (NIR-CI) is widely used in the pharmaceutical industry not only to provide the concentration of a compound of interest but far more often to obtain the spatial distribution of different ingredients within the considered sample like a single tablet. For such sample characterization, having a high field of view is of major interest. As well as their high field of view, recent NIR-CI spectrometers have the ability to acquire thousands of spectra in a very short time due to focal plane array detector they use. Nevertheless, the spatial resolution is often limited which implies that generated chemical images are often biased. In view of this it was deemed appropriate to develop a new chemometrics methodology called “super-resolution” in order to increase the spatial resolution of spectroscopic images. The main idea is the fusion of several low-resolution images of the same sample observed from different point of view in order to generate one higher-resolution image. We offer here an objective and quantitative evaluation of the super-resolution concept with applications on pharmaceutical solid samples.

Performance Analysis of Wavelet Transforms for Learning based Single frame Image Super-resolution

super resolution concept has been introduced for image enhancement in various applications. Image enhancement is crucial operation essential for reducing different possible degradations of the captured image. More sophisticated techniques are already proposed. Wavelet transform based algorithms are widely used in many applications. Wavelet transform/s is used to extrapolate missing high frequency components which improve the efficiency of an algorithm. In this paper for super resolving the images, wavelet coefficients of the unknown high resolution image are learnt from a set of high resolution training images in wavelet domain. The performance of different discrete orthogonal and a biorthogonal wavelets have been evaluated on different class of images in terms of MSE and PSNR. The outcome of this work suggests that use of db4 wavelet transform is appropriate for super resolution technique. The PSNR obtained with this transform outfits for other wavelet transforms.

Spot Variation Fluorescence Correlation Spectroscopy Allows for Superresolution Chronoscopy of Confinement Times in Membranes

Resolving the dynamical interplay of proteins and lipids in the live-cell plasma membrane represents a central goal in current cell biology. Superresolution concepts have introduced a means of capturing spatial heterogeneity at a nanoscopic length scale. Similar concepts for detecting dynamical transitions (superresolution chronoscopy) are still lacking. Here, we show that recently introduced spot-variation fluorescence correlation spectroscopy allows for sensing transient confinement times of membrane constituents at dramatically improved resolution. Using standard diffraction-limited optics, spot-variation fluorescence correlation spectroscopy captures signatures of single retardation events far below the transit time of the tracer through the focal spot. We provide an analytical description of special cases of transient binding of a tracer to pointlike traps, or association of a tracer with nanodomains. The influence of trap mobility and the underlying binding kinetics are quantified. Experimental approaches are suggested that allow for gaining quantitative mechanistic insights into the interaction processes of membrane constituents.

Infrared chemical imaging: Spatial resolution evaluation and super-resolution concept

Chemical imaging systems help to solve many challenges in various scientific fields. Able to deliver rapid spatial and chemical information, modern infrared spectrometers using Focal Plane Array detectors (FPA) are of great interest. Considering conventional infrared spectrometers with a single element detector, we can consider that the diffraction-limited spatial resolution is more or less equal to the wavelength of the light (i.e. 2.5–25 μm). Unfortunately, the spatial resolution of FPA spectroscopic setup is even lower due to the detector pixel size. This becomes a real constraint when micron-sized samples are analysed. New chemometrics methods are thus of great interest to overcome such resolution drawback, while keeping our far-field infrared imaging spectrometers. The aim of the present work is to evaluate the super-resolution concept in order to increase the spatial resolution of infrared imaging spectrometers using FPA detectors. The main idea of super-resolution is the fusion of several low-resolution images of the same sample to obtain a higher-resolution image. Applying the super-resolution concept on a relatively low number of FPA acquisitions, it was possible to observe a 30% decrease in spatial resolution.

Super-resolution, degrees of freedom and synthetic aperture radar

Super-resolution concepts offer the potential of resolution beyond the classical limit. This great promise has not generally been realised. The potential application of super-resolution concepts to synthetic aperture radar is investigated. The analytical basis for super-resolution theory is discussed. The application of the concept to synthetic aperture radar is investigated as an operator inversion problem. Generally, the operator inversion problem is ill posed. However, it is shown that it is the particular form of the eigenvalue spectrum of the associated operator that precludes any significant resolution enhancement of synthetic aperture radar images. A criterion for judging super-resolution processing of an image is presented.

Time multiplexing superresolution based on interference grating projection.

In a previous work done by the authors, it was shown that the superresolution concept based on two moving gratings could be effected by a physical grating attached to the object and a virtual grating. This concept was shown to be very efficient and exhibited features that are helpful in removing some artifacts caused when coherent illumination is used. Furthermore, it simplifies the optical and mechanical modules of the super-resolving system by removing the need for mechanical movement of one grating. However, the system still required the need for moving the first (encoding) grating attached to the input. In this study the encoding grating is replaced by use of a projected grating. This approach simplifies the need for attaching the grating to the input object and thus new applications, such as remote sensing can be considered. The theoretical concept is demonstrated and experimental results are shown.