Showing papers on "Image quality published in 2018"
TL;DR: This paper provides a deep learning-based strategy for reconstruction of CS-MRI, and bridges a substantial gap between conventional non-learning methods working only on data from a single image, and prior knowledge from large training data sets.
Abstract: Compressed sensing magnetic resonance imaging (CS-MRI) enables fast acquisition, which is highly desirable for numerous clinical applications. This can not only reduce the scanning cost and ease patient burden, but also potentially reduce motion artefacts and the effect of contrast washout, thus yielding better image quality. Different from parallel imaging-based fast MRI, which utilizes multiple coils to simultaneously receive MR signals, CS-MRI breaks the Nyquist–Shannon sampling barrier to reconstruct MRI images with much less required raw data. This paper provides a deep learning-based strategy for reconstruction of CS-MRI, and bridges a substantial gap between conventional non-learning methods working only on data from a single image, and prior knowledge from large training data sets. In particular, a novel conditional Generative Adversarial Networks-based model (DAGAN)-based model is proposed to reconstruct CS-MRI. In our DAGAN architecture, we have designed a refinement learning method to stabilize our U-Net based generator, which provides an end-to-end network to reduce aliasing artefacts. To better preserve texture and edges in the reconstruction, we have coupled the adversarial loss with an innovative content loss. In addition, we incorporate frequency-domain information to enforce similarity in both the image and frequency domains. We have performed comprehensive comparison studies with both conventional CS-MRI reconstruction methods and newly investigated deep learning approaches. Compared with these methods, our DAGAN method provides superior reconstruction with preserved perceptual image details. Furthermore, each image is reconstructed in about 5 ms, which is suitable for real-time processing.
835 citations
TL;DR: The proposed approach relies on the success (and retraining) of proven, state-of-the-art deep object recognition networks and can be used to not only score images reliably and with high correlation to human perception, but also to assist with adaptation and optimization of photo editing/enhancement algorithms in a photographic pipeline.
Abstract: Automatically learned quality assessment for images has recently become a hot topic due to its usefulness in a wide variety of applications, such as evaluating image capture pipelines, storage techniques, and sharing media. Despite the subjective nature of this problem, most existing methods only predict the mean opinion score provided by data sets, such as AVA and TID2013. Our approach differs from others in that we predict the distribution of human opinion scores using a convolutional neural network. Our architecture also has the advantage of being significantly simpler than other methods with comparable performance. Our proposed approach relies on the success (and retraining) of proven, state-of-the-art deep object recognition networks. Our resulting network can be used to not only score images reliably and with high correlation to human perception, but also to assist with adaptation and optimization of photo editing/enhancement algorithms in a photographic pipeline. All this is done without need for a “golden” reference image, consequently allowing for single-image, semantic- and perceptually-aware, no-reference quality assessment.
606 citations
TL;DR: This work introduces an effective technique to enhance the images captured underwater and degraded due to the medium scattering and absorption by building on the blending of two images that are directly derived from a color-compensated and white-balanced version of the original degraded image.
Abstract: We introduce an effective technique to enhance the images captured underwater and degraded due to the medium scattering and absorption. Our method is a single image approach that does not require specialized hardware or knowledge about the underwater conditions or scene structure. It builds on the blending of two images that are directly derived from a color-compensated and white-balanced version of the original degraded image. The two images to fusion, as well as their associated weight maps, are defined to promote the transfer of edges and color contrast to the output image. To avoid that the sharp weight map transitions create artifacts in the low frequency components of the reconstructed image, we also adapt a multiscale fusion strategy. Our extensive qualitative and quantitative evaluation reveals that our enhanced images and videos are characterized by better exposedness of the dark regions, improved global contrast, and edges sharpness. Our validation also proves that our algorithm is reasonably independent of the camera settings, and improves the accuracy of several image processing applications, such as image segmentation and keypoint matching.
601 citations
TL;DR: A deep neural network-based approach to image quality assessment (IQA) that allows for joint learning of local quality and local weights in an unified framework and shows a high ability to generalize between different databases, indicating a high robustness of the learned features.
Abstract: We present a deep neural network-based approach to image quality assessment (IQA). The network is trained end-to-end and comprises ten convolutional layers and five pooling layers for feature extraction, and two fully connected layers for regression, which makes it significantly deeper than related IQA models. Unique features of the proposed architecture are that: 1) with slight adaptations it can be used in a no-reference (NR) as well as in a full-reference (FR) IQA setting and 2) it allows for joint learning of local quality and local weights, i.e., relative importance of local quality to the global quality estimate, in an unified framework. Our approach is purely data-driven and does not rely on hand-crafted features or other types of prior domain knowledge about the human visual system or image statistics. We evaluate the proposed approach on the LIVE, CISQ, and TID2013 databases as well as the LIVE In the wild image quality challenge database and show superior performance to state-of-the-art NR and FR IQA methods. Finally, cross-database evaluation shows a high ability to generalize between different databases, indicating a high robustness of the learned features.
479 citations
TL;DR: RefineGAN as mentioned in this paper is a variant of fully-residual convolutional autoencoder and generative adversarial networks (GANs) specifically designed for CS-MRI formulation; it employs deeper generator and discriminator networks with cyclic data consistency loss for faithful interpolation in the given under-sampled data.
Abstract: Compressed sensing magnetic resonance imaging (CS-MRI) has provided theoretical foundations upon which the time-consuming MRI acquisition process can be accelerated. However, it primarily relies on iterative numerical solvers, which still hinders their adaptation in time-critical applications. In addition, recent advances in deep neural networks have shown their potential in computer vision and image processing, but their adaptation to MRI reconstruction is still in an early stage. In this paper, we propose a novel deep learning-based generative adversarial model, RefineGAN , for fast and accurate CS-MRI reconstruction. The proposed model is a variant of fully-residual convolutional autoencoder and generative adversarial networks (GANs), specifically designed for CS-MRI formulation; it employs deeper generator and discriminator networks with cyclic data consistency loss for faithful interpolation in the given under-sampled $k$ -space data. In addition, our solution leverages a chained network to further enhance the reconstruction quality. RefineGAN is fast and accurate—the reconstruction process is extremely rapid, as low as tens of milliseconds for reconstruction of a $256\times 256$ image, because it is one-way deployment on a feed-forward network, and the image quality is superior even for extremely low sampling rate (as low as 10%) due to the data-driven nature of the method. We demonstrate that RefineGAN outperforms the state-of-the-art CS-MRI methods by a large margin in terms of both running time and image quality via evaluation using several open-source MRI databases.
428 citations
TL;DR: This work demonstrates the strong competitiveness of MEON against state-of-the-art BIQA models using the group maximum differentiation competition methodology and empirically demonstrates that GDN is effective at reducing model parameters/layers while achieving similar quality prediction performance.
Abstract: We propose a multi-task end-to-end optimized deep neural network (MEON) for blind image quality assessment (BIQA). MEON consists of two sub-networks—a distortion identification network and a quality prediction network—sharing the early layers. Unlike traditional methods used for training multi-task networks, our training process is performed in two steps. In the first step, we train a distortion type identification sub-network, for which large-scale training samples are readily available. In the second step, starting from the pre-trained early layers and the outputs of the first sub-network, we train a quality prediction sub-network using a variant of the stochastic gradient descent method. Different from most deep neural networks, we choose biologically inspired generalized divisive normalization (GDN) instead of rectified linear unit as the activation function. We empirically demonstrate that GDN is effective at reducing model parameters/layers while achieving similar quality prediction performance. With modest model complexity, the proposed MEON index achieves state-of-the-art performance on four publicly available benchmarks. Moreover, we demonstrate the strong competitiveness of MEON against state-of-the-art BIQA models using the group maximum differentiation competition methodology.
391 citations
20 Jan 2018
TL;DR: In this article, a diffuser placed in front of an image sensor is used for single-shot 3D imaging, which exploits sparsity in the sample to solve for more 3D voxels than pixels on the 2D sensor.
Abstract: We demonstrate a compact, easy-to-build computational camera for single-shot three-dimensional (3D) imaging. Our lensless system consists solely of a diffuser placed in front of an image sensor. Every point within the volumetric field-of-view projects a unique pseudorandom pattern of caustics on the sensor. By using a physical approximation and simple calibration scheme, we solve the large-scale inverse problem in a computationally efficient way. The caustic patterns enable compressed sensing, which exploits sparsity in the sample to solve for more 3D voxels than pixels on the 2D sensor. Our 3D reconstruction grid is chosen to match the experimentally measured two-point optical resolution, resulting in 100 million voxels being reconstructed from a single 1.3 megapixel image. However, the effective resolution varies significantly with scene content. Because this effect is common to a wide range of computational cameras, we provide a new theory for analyzing resolution in such systems.
369 citations
18 Jun 2018
TL;DR: This work proposes a novel hierarchical approach for text-to-image synthesis by inferring semantic layout and shows that the model can substantially improve the image quality, interpretability of output and semantic alignment to input text over existing approaches.
Abstract: We propose a novel hierarchical approach for text-to-image synthesis by inferring semantic layout. Instead of learning a direct mapping from text to image, our algorithm decomposes the generation process into multiple steps, in which it first constructs a semantic layout from the text by the layout generator and converts the layout to an image by the image generator. The proposed layout generator progressively constructs a semantic layout in a coarse-to-fine manner by generating object bounding boxes and refining each box by estimating object shapes inside the box. The image generator synthesizes an image conditioned on the inferred semantic layout, which provides a useful semantic structure of an image matching with the text description. Our model not only generates semantically more meaningful images, but also allows automatic annotation of generated images and user-controlled generation process by modifying the generated scene layout. We demonstrate the capability of the proposed model on challenging MS-COCO dataset and show that the model can substantially improve the image quality, interpretability of output and semantic alignment to input text over existing approaches.
360 citations
University of Tokyo1, Princeton University2, University of Geneva3, National Institutes of Natural Sciences, Japan4, Graduate University for Advanced Studies5, University of Toronto6, California Institute of Technology7, Hiroshima University8, Academia Sinica Institute of Astronomy and Astrophysics9, Tohoku University10, National Tsing Hua University11, Polish Academy of Sciences12, National Taiwan Normal University13, University of California, Santa Cruz14, Carnegie Mellon University15, Ehime University16, Kobe University17, Lawrence Berkeley National Laboratory18, University of California, Berkeley19, Nagoya University20, University of California, Riverside21, Harvard University22, National Central University23, Mahidol University24
TL;DR: This paper presents the second data release of the Hyper Suprime-Cam Subaru Strategic Program, a wide-field optical imaging survey on the 8.2 meter Subaru Telescope, including a major update to the processing pipeline, including improved sky subtraction, PSF modeling, object detection, and artifact rejection.
Abstract: This paper presents the second data release of the Hyper Suprime-Cam Subaru Strategic Program, a wide-field optical imaging survey using the 8.2 m Subaru Telescope. The release includes data from 174 nights of observation through 2018 January. The Wide layer data cover about 300 deg|$^2$| in all five broad-band filters (|$grizy$|) to the nominal survey exposure (10 min in |$gr$| and 20 min in |$izy$|). Partially observed areas are also included in the release; about 1100 deg|$^2$| is observed in at least one filter and one exposure. The median seeing in the i-band is |${0_{.}^{\prime \prime }6}$|, demonstrating the superb image quality of the survey. The Deep (26 deg|$^2$|) and UltraDeep (4 deg|$^2$|) data are jointly processed and the UltraDeep-COSMOS field reaches an unprecedented depth of |$i\sim 28$| at |$5 \, \sigma$| for point sources. In addition to the broad-band data, narrow-band data are also available in the Deep and UltraDeep fields. This release includes a major update to the processing pipeline, including improved sky subtraction, PSF modeling, object detection, and artifact rejection. The overall data quality has been improved, but this release is not without problems; there is a persistent deblender problem as well as new issues with masks around bright stars. The user is encouraged to review the issue list before utilizing the data for scientific explorations. All the image products as well as catalog products are available for download. The catalogs are also loaded into a database, which provides an easy interface for users to retrieve data for objects of interest. In addition to these main data products, detailed galaxy shape measurements withheld from Public Data Release 1 (PDR1) are now available to the community. The shape catalog is drawn from the S16A internal release, which has a larger area than PDR1 (160 deg|$^2$|). All products are available at the data release site, https://hsc-release.mtk.nao.ac.jp/.
348 citations
TL;DR: A new no-reference (NR) IQA model is developed and a robust image enhancement framework is established based on quality optimization, which can well enhance natural images, low-contrast images,Low-light images, and dehazed images.
Abstract: In this paper, we investigate into the problem of image quality assessment (IQA) and enhancement via machine learning. This issue has long attracted a wide range of attention in computational intelligence and image processing communities, since, for many practical applications, e.g., object detection and recognition, raw images are usually needed to be appropriately enhanced to raise the visual quality (e.g., visibility and contrast). In fact, proper enhancement can noticeably improve the quality of input images, even better than originally captured images, which are generally thought to be of the best quality. In this paper, we present two most important contributions. The first contribution is to develop a new no-reference (NR) IQA model. Given an image, our quality measure first extracts 17 features through analysis of contrast, sharpness, brightness and more, and then yields a measure of visual quality using a regression module, which is learned with big-data training samples that are much bigger than the size of relevant image data sets. The results of experiments on nine data sets validate the superiority and efficiency of our blind metric compared with typical state-of-the-art full-reference, reduced-reference and NA IQA methods. The second contribution is that a robust image enhancement framework is established based on quality optimization. For an input image, by the guidance of the proposed NR-IQA measure, we conduct histogram modification to successively rectify image brightness and contrast to a proper level. Thorough tests demonstrate that our framework can well enhance natural images, low-contrast images, low-light images, and dehazed images. The source code will be released at https://sites.google.com/site/guke198701/publications .
297 citations
TL;DR: Experimental results show that the proposed 3D c‐GANs method outperforms the benchmark methods and achieves much better performance than the state‐of‐the‐art methods in both qualitative and quantitative measures.
TL;DR: A trainable Convolutional Neural Network is proposed for weakly illuminated image enhancement, namely LightenNet, which takes a weakly illumination image as input and outputs its illumination map that is subsequently used to obtain the enhanced image based on Retinex model.
TL;DR: This work presents a new deep learning approach to blending for IBR, in which held-out real image data is used to learn blending weights to combine input photo contributions, and designs the network architecture and the training loss to provide high quality novel view synthesis, while reducing temporal flickering artifacts.
Abstract: Free-viewpoint image-based rendering (IBR) is a standing challenge. IBR methods combine warped versions of input photos to synthesize a novel view. The image quality of this combination is directly affected by geometric inaccuracies of multi-view stereo (MVS) reconstruction and by view- and image-dependent effects that produce artifacts when contributions from different input views are blended. We present a new deep learning approach to blending for IBR, in which we use held-out real image data to learn blending weights to combine input photo contributions. Our Deep Blending method requires us to address several challenges to achieve our goal of interactive free-viewpoint IBR navigation. We first need to provide sufficiently accurate geometry so the Convolutional Neural Network (CNN) can succeed in finding correct blending weights. We do this by combining two different MVS reconstructions with complementary accuracy vs. completeness tradeoffs. To tightly integrate learning in an interactive IBR system, we need to adapt our rendering algorithm to produce a fixed number of input layers that can then be blended by the CNN. We generate training data with a variety of captured scenes, using each input photo as ground truth in a held-out approach. We also design the network architecture and the training loss to provide high quality novel view synthesis, while reducing temporal flickering artifacts. Our results demonstrate free-viewpoint IBR in a wide variety of scenes, clearly surpassing previous methods in visual quality, especially when moving far from the input cameras.
TL;DR: The best proposal, named DeepBIQ, estimates the image quality by average-pooling the scores predicted on multiple subregions of the original image, having a linear correlation coefficient with human subjective scores of almost 0.91.
Abstract: In this work, we investigate the use of deep learning for distortion-generic blind image quality assessment. We report on different design choices, ranging from the use of features extracted from pre-trained convolutional neural networks (CNNs) as a generic image description, to the use of features extracted from a CNN fine-tuned for the image quality task. Our best proposal, named DeepBIQ, estimates the image quality by average-pooling the scores predicted on multiple subregions of the original image. Experimental results on the LIVE In the Wild Image Quality Challenge Database show that DeepBIQ outperforms the state-of-the-art methods compared, having a linear correlation coefficient with human subjective scores of almost 0.91. These results are further confirmed also on four benchmark databases of synthetically distorted images: LIVE, CSIQ, TID2008, and TID2013.
TL;DR: Data quality both inflated and obscured associations with age during adolescence, indicating that reliable measures of data quality can be automatically derived from T1‐weighted volumes, and that failing to control for dataquality can systematically bias the results of studies of brain maturation.
TL;DR: To develop a super‐resolution technique using convolutional neural networks for generating thin‐slice knee MR images from thicker input slices, and compare this method with alternative through‐plane interpolation methods.
Abstract: PURPOSE To develop a super-resolution technique using convolutional neural networks for generating thin-slice knee MR images from thicker input slices, and compare this method with alternative through-plane interpolation methods. METHODS We implemented a 3D convolutional neural network entitled DeepResolve to learn residual-based transformations between high-resolution thin-slice images and lower-resolution thick-slice images at the same center locations. DeepResolve was trained using 124 double echo in steady-state (DESS) data sets with 0.7-mm slice thickness and tested on 17 patients. Ground-truth images were compared with DeepResolve, clinically used tricubic interpolation, and Fourier interpolation methods, along with state-of-the-art single-image sparse-coding super-resolution. Comparisons were performed using structural similarity, peak SNR, and RMS error image quality metrics for a multitude of thin-slice downsampling factors. Two musculoskeletal radiologists ranked the 3 data sets and reviewed the diagnostic quality of the DeepResolve, tricubic interpolation, and ground-truth images for sharpness, contrast, artifacts, SNR, and overall diagnostic quality. Mann-Whitney U tests evaluated differences among the quantitative image metrics, reader scores, and rankings. Cohen's Kappa (κ) evaluated interreader reliability. RESULTS DeepResolve had significantly better structural similarity, peak SNR, and RMS error than tricubic interpolation, Fourier interpolation, and sparse-coding super-resolution for all downsampling factors (p < .05, except 4 × and 8 × sparse-coding super-resolution downsampling factors). In the reader study, DeepResolve significantly outperformed (p < .01) tricubic interpolation in all image quality categories and overall image ranking. Both readers had substantial scoring agreement (κ = 0.73). CONCLUSION DeepResolve was capable of resolving high-resolution thin-slice knee MRI from lower-resolution thicker slices, achieving superior quantitative and qualitative diagnostic performance to both conventionally used and state-of-the-art methods.
TL;DR: NanoJ-SQUIRREL is presented, an ImageJ-based analytical approach that provides quantitative assessment of super-resolution image quality by comparing diffraction-limited images and super- resolution equivalents of the same acquisition volume, and can guide researchers in optimizing imaging parameters.
Abstract: Super-resolution microscopy depends on steps that can contribute to the formation of image artifacts, leading to misinterpretation of biological information. We present NanoJ-SQUIRREL, an ImageJ-based analytical approach that provides quantitative assessment of super-resolution image quality. By comparing diffraction-limited images and super-resolution equivalents of the same acquisition volume, this approach generates a quantitative map of super-resolution defects and can guide researchers in optimizing imaging parameters.
TL;DR: Comparative studies on five large IQA databases show that the proposed BPRI model is comparable to the state-of-the-art opinion-aware- and OU-BIQA models, and not only performs well on natural scene images, but also is applicable to screen content images.
Abstract: Traditional full-reference image quality assessment (IQA) metrics generally predict the quality of the distorted image by measuring its deviation from a perfect quality image called reference image. When the reference image is not fully available, the reduced-reference and no-reference IQA metrics may still be able to derive some characteristics of the perfect quality images, and then measure the distorted image's deviation from these characteristics. In this paper, contrary to the conventional IQA metrics, we utilize a new “reference” called pseudo-reference image (PRI) and a PRI-based blind IQA (BIQA) framework. Different from a traditional reference image, which is assumed to have a perfect quality, PRI is generated from the distorted image and is assumed to suffer from the severest distortion for a given application. Based on the PRI-based BIQA framework, we develop distortion-specific metrics to estimate blockiness, sharpness, and noisiness. The PRI-based metrics calculate the similarity between the distorted image's and the PRI's structures. An image suffering from severer distortion has a higher degree of similarity with the corresponding PRI. Through a two-stage quality regression after a distortion identification framework, we then integrate the PRI-based distortion-specific metrics into a general-purpose BIQA method named blind PRI-based (BPRI) metric. The BPRI metric is opinion-unaware (OU) and almost training-free except for the distortion identification process. Comparative studies on five large IQA databases show that the proposed BPRI model is comparable to the state-of-the-art opinion-aware- and OU-BIQA models. Furthermore, BPRI not only performs well on natural scene images, but also is applicable to screen content images. The MATLAB source code of BPRI and other PRI-based distortion-specific metrics will be publicly available.
TL;DR: This paper introduces multiple pseudo reference images (MPRIs) by further degrading the distorted image in several ways and to certain degrees, and then compares the similarities between the distorted images and the MPRIs, and uses the full-reference IQA framework to compute the quality.
Abstract: Traditional blind image quality assessment (IQA) measures generally predict quality from a sole distorted image directly. In this paper, we first introduce multiple pseudo reference images (MPRIs) by further degrading the distorted image in several ways and to certain degrees, and then compare the similarities between the distorted image and the MPRIs. Via such distortion aggravation, we can have some references to compare with, i.e., the MPRIs, and utilize the full-reference IQA framework to compute the quality. Specifically, we apply four types and five levels of distortion aggravation to deal with the commonly encountered distortions. Local binary pattern features are extracted to describe the similarities between the distorted image and the MPRIs. The similarity scores are then utilized to estimate the overall quality. More similar to a specific pseudo reference image (PRI) indicates closer quality to this PRI. Owning to the availability of the created multiple PRIs, we can reduce the influence of image content, and infer the image quality more accurately and consistently. Validation is conducted on four mainstream natural scene image and screen content image quality assessment databases, and the proposed method is comparable to or outperforms the state-of-the-art blind IQA measures. The MATLAB source code of the proposed measure will be publicly available.
National Institutes of Natural Sciences, Japan1, Graduate University for Advanced Studies2, Hiroshima University3, University of Tokyo4, Institute for the Physics and Mathematics of the Universe5, Princeton University6, Jet Propulsion Laboratory7, Nagoya University8, Tohoku University9, Kyoto Sangyo University10, Academia Sinica Institute of Astronomy and Astrophysics11, Chungshan Institute of Science and Technology12, Hamamatsu Photonics13, Mitsubishi Electric14
20 Apr 2018
TL;DR: Using a table-top x-ray source, ghost imaging of plane and natural objects with ultra-low radiation on the order of single photons is realized and a higher contrast-to-noise ratio is obtained for the same radiation dose.
Abstract: Computational ghost imaging, in which an image is retrieved from a known patterned field that illuminates an object and the total transmitted intensity therefrom, has seen great advances on account of its advantages and potential applications at all wavelengths. However, even though lensless x-ray ghost imaging was anticipated more than a decade ago, its development has been hampered due to the lack of suitable optics. The image quality is proportional to the total flux in conventional projection x-ray imaging, but high photon energy could severely damage the object being imaged, so decreasing the radiation dose while maintaining image quality is a fundamental problem. Using a simple tabletop x-ray source, we have successfully realized ghost imaging of planar and natural objects with a much higher contrast-to-noise ratio compared to projection x-ray imaging at the same low-radiation dose. Ultra-low-flux imaging has been achieved, and thus radiation damage of biological specimens could be greatly reduced with this new technique.
TL;DR: A real-time fitter for 3D single-molecule localization microscopy using experimental point spread functions (PSFs) that achieves minimal uncertainty in 3D on any microscope and is compatible with any PSF engineering approach is presented.
Abstract: We present a real-time fitter for 3D single-molecule localization microscopy using experimental point spread functions (PSFs) that achieves minimal uncertainty in 3D on any microscope and is compatible with any PSF engineering approach. We used this method to image cellular structures and attained unprecedented image quality for astigmatic PSFs. The fitter compensates for most optical aberrations and makes accurate 3D super-resolution microscopy broadly accessible, even on standard microscopes without dedicated 3D optics.
TL;DR: This paper proposes a novel 3-D noise reduction method, called structurally sensitive multi-scale generative adversarial net, to improve the low-dose CT image quality, which incorporates3-D volumetric information to improved the image quality.
Abstract: Computed tomography (CT) is a popular medical imaging modality and enjoys wide clinical applications. At the same time, the X-ray radiation dose associated with CT scannings raises a public concern due to its potential risks to the patients. Over the past years, major efforts have been dedicated to the development of low-dose CT (LDCT) methods. However, the radiation dose reduction compromises the signal-to-noise ratio, leading to strong noise and artifacts that down-grade the CT image quality. In this paper, we propose a novel 3-D noise reduction method, called structurally sensitive multi-scale generative adversarial net, to improve the LDCT image quality. Specifically, we incorporate 3-D volumetric information to improve the image quality. Also, different loss functions for training denoising models are investigated. Experiments show that the proposed method can effectively preserve the structural and textural information in reference to the normal-dose CT images and significantly suppress noise and artifacts. Qualitative visual assessments by three experienced radiologists demonstrate that the proposed method retrieves more information and outperforms competing methods.
TL;DR: The Haar wavelet-based perceptual similarity index (HaarPSI) as discussed by the authors was proposed to assess local similarities between two images, as well as the relative importance of image areas.
Abstract: In most practical situations, the compression or transmission of images and videos creates distortions that will eventually be perceived by a human observer. Vice versa, image and video restoration techniques, such as inpainting or denoising, aim to enhance the quality of experience of human viewers. Correctly assessing the similarity between an image and an undistorted reference image as subjectively experienced by a human viewer can thus lead to significant improvements in any transmission, compression, or restoration system. This paper introduces the Haar wavelet-based perceptual similarity index (HaarPSI), a novel and computationally inexpensive similarity measure for full reference image quality assessment. The HaarPSI utilizes the coefficients obtained from a Haar wavelet decomposition to assess local similarities between two images, as well as the relative importance of image areas. The consistency of the HaarPSI with the human quality of experience was validated on four large benchmark databases containing thousands of differently distorted images. On these databases, the HaarPSI achieves higher correlations with human opinion scores than state-of-the-art full reference similarity measures like the structural similarity index (SSIM), the feature similarity index (FSIM), and the visual saliency-based index (VSI). Along with the simple computational structure and the short execution time, these experimental results suggest a high applicability of the HaarPSI in real world tasks.
08 Sep 2018
TL;DR: MSG-Net is the first to achieve real-time brush-size control in a purely feed-forward manner for style transfer and is compatible with most existing techniques including content-style interpolation, color-preserving, spatial control and brush stroke size control.
Abstract: Despite the rapid progress in style transfer, existing approaches using feed-forward generative network for multi-style or arbitrary-style transfer are usually compromised of image quality and model flexibility. We find it is fundamentally difficult to achieve comprehensive style modeling using 1-dimensional style embedding. Motivated by this, we introduce CoMatch Layer that learns to match the second order feature statistics with the target styles. With the CoMatch Layer, we build a Multi-style Generative Network (MSG-Net), which achieves real-time performance. In addition, we employ an specific strategy of upsampled convolution which avoids checkerboard artifacts caused by fractionally-strided convolution. Our method has achieved superior image quality comparing to state-of-the-art approaches. The proposed MSG-Net as a general approach for real-time style transfer is compatible with most existing techniques including content-style interpolation, color-preserving, spatial control and brush stroke size control. MSG-Net is the first to achieve real-time brush-size control in a purely feed-forward manner for style transfer. Our implementations and pre-trained models for Torch, PyTorch and MXNet frameworks will be publicly available (Links can be found at http://hangzhang.org/).
TL;DR: This work proposes a novel codebook-based BIQA method by optimizing multistage discriminative dictionaries (MSDDs), which has been evaluated on five databases and experimental results well confirm its superiority over existing relevant BIZA methods.
Abstract: State-of-the-art algorithms for blind image quality assessment (BIQA) typically have two categories. The first category approaches extract natural scene statistics (NSS) as features based on the statistical regularity of natural images. The second category approaches extract features by feature encoding with respect to a learned codebook. However, several problems need to be addressed in existing codebook-based BIQA methods. First, the high-dimensional codebook-based features are memory-consuming and have the risk of over-fitting. Second, there is a semantic gap between the constructed codebook by unsupervised learning and image quality. To address these problems, we propose a novel codebook-based BIQA method by optimizing multistage discriminative dictionaries (MSDDs). To be specific, MSDDs are learned by performing the label consistent K-SVD (LC-KSVD) algorithm in a stage-by-stage manner. For each stage, a new quality consistency constraint called “quality-discriminative regularization” term is introduced and incorporated into the reconstruction error term to form a unified objective function, which can be effectively solved by LC-KSVD for discriminative dictionary learning. Then, the latter stage takes the reconstruction residual data in the former stage as input based on which LC-KSVD is repeatedly performed until the final stage is reached. Once the MSDDs are learned, multistage feature encoding is performed to extract feature codes. Finally, the feature codes are concatenated across all stages and aggregated over the entire image for quality prediction via regression. The proposed method has been evaluated on five databases and experimental results well confirm its superiority over existing relevant BIQA methods.
TL;DR: A computational ghost imaging scheme, which utilizes an LED-based, high-speed illumination module is presented, which provides a cost-effective and high- speed imaging technique for dynamic imaging applications.
Abstract: Single-pixel imaging uses a single-pixel detector, rather than a focal plane detector array, to image a scene. It provides advantages for applications such as multi-wavelength, three-dimensional imaging. However, low frame rates have been a major obstacle inhibiting the use of computational ghost imaging technique in wider applications since its invention one decade ago. To address this problem, a computational ghost imaging scheme, which utilizes an LED-based, high-speed illumination module is presented in this work. At 32 × 32 pixel resolution, the proof-of-principle system achieved continuous imaging with 1000 fps frame rate, approximately two orders larger than those of other existing ghost imaging systems. The proposed scheme provides a cost-effective and high-speed imaging technique for dynamic imaging applications.
TL;DR: An MRI reconstruction algorithm, which is referred to as quantitative susceptibility mapping (QSM), has been developed using a deep neural network in order to perform dipole deconvolution, which restores magnetic susceptibility source from an MRI field map.
24 Sep 2018
TL;DR: A new dataset -named I-HAZE- that contains 35 image pairs of hazy and corresponding haze-free (ground-truth) indoor images that allows us to objectively compare the existing image dehazing techniques using traditional image quality metrics such as PSNR and SSIM.
Abstract: Image dehazing has become an important computational imaging topic in the recent years. However, due to the lack of ground truth images, the comparison of dehazing methods is not straightforward, nor objective. To overcome this issue we introduce I-HAZE, a new dataset that contains 35 image pairs of hazy and corresponding haze-free (ground-truth) indoor images. Different from most of the existing dehazing databases, hazy images have been generated using real haze produced by a professional haze machine. To ease color calibration and improve the assessment of dehazing algorithms, each scene includes a MacBeth color checker. Moreover, since the images are captured in a controlled environment, both haze-free and hazy images are captured under the same illumination conditions. This represents an important advantage of the I-HAZE dataset that allows us to objectively compare the existing image dehazing techniques using traditional image quality metrics such as PSNR and SSIM.
TL;DR: It is suggested that SRCNN may become a potential solution for generating high-resolution CT images from standard CT images and significantly outperforms the linear interpolation methods for enhancing image resolution in chest CT images.
Abstract: In this study, the super-resolution convolutional neural network (SRCNN) scheme, which is the emerging deep-learning-based super-resolution method for enhancing image resolution in chest CT images, was applied and evaluated using the post-processing approach. For evaluation, 89 chest CT cases were sampled from The Cancer Imaging Archive. The 89 CT cases were divided randomly into 45 training cases and 44 external test cases. The SRCNN was trained using the training dataset. With the trained SRCNN, a high-resolution image was reconstructed from a low-resolution image, which was down-sampled from an original test image. For quantitative evaluation, two image quality metrics were measured and compared to those of the conventional linear interpolation methods. The image restoration quality of the SRCNN scheme was significantly higher than that of the linear interpolation methods (p < 0.001 or p < 0.05). The high-resolution image reconstructed by the SRCNN scheme was highly restored and comparable to the original reference image, in particular, for a ×2 magnification. These results indicate that the SRCNN scheme significantly outperforms the linear interpolation methods for enhancing image resolution in chest CT images. The results also suggest that SRCNN may become a potential solution for generating high-resolution CT images from standard CT images.