Search

Ultra high resolution (UHR) images are almost always downsampled to fit small displays of mobile end devices and upsampled to its original resolution when exhibited on very high-resolution displays. This observation motivates us on jointly optimizing operation pairs of downsampling and upsampling that are spatially adaptive to image contents for maximal rate-distortion performance. In this paper, we propose an adaptive downsampled dual-layer (ADDL) image compression system. In the ADDL compression system, an image is reduced in resolution by learned content-adaptive downsampling kernels and compressed to form a coded base layer. For decompression the base layer is decoded and upconverted to the original resolution using a deep upsampling neural network, aided by the prior knowledge of the learned adaptive downsampling kernels. We restrict the downsampling kernels to the form of Gabor filters in order to reduce the complexity of filter optimization and also reduce the amount of side information needed by the decoder for adaptive upsampling. Extensive experiments demonstrate that the proposed ADDL compression approach of jointly optimized, spatially adaptive downsampling and upconversion outperforms the state of the art image compression methods.

There is no effective treatment for peripheral nerve injury-induced chronic neuropathic pain (NP), which profoundly impacts the quality of life of those affected. Transmembraneprotein100 (TMEM100) is considered to be a pain regulatory protein and is expressed in the dorsal root ganglion (DRG) of rats. However, the mechanism of pain regulation and the expression of TMEM100 following various peripheral nerve injuries are unclear. In this study, we constructed two pain models of peripheral nerve injury: tibial nerve injury (TNI) and chronic constriction injury (CCI). This study found that the Paw Withdrawal Mechanical Threshold (PWMT) and Paw Withdraw Thermal Latency (PWTL) of the rats in the two pain models decreased significantly, and the expression of TMEM100 in the DRG of two groups also decreased significantly. Furthermore, the decrease in the CCI group was more obvious than in the TNI group. There was no significant statistical significance (P>0.05). We constructed an adeno-associated virus 6 (AAV6) vector expressing recombinant fluorescent TMEM100 protein and injected it into the sciatic nerve (SN) of two pain models: CCI and TNI. PWMT and PWTL were significantly increased in the two groups, along with the expression of TMEM100 in the spinal cord and DRG. It also significantly inhibited the activation of microglia, astrocytes, and several inflammatory mediators (TNF- α, IL-1 β, and IL-6). In summary, the results of this study suggested that TMEM100 might be a promising molecular strategy for the treatment of NP, and its anti-inflammatory effects might play an important role in pain relief.

Recent supervised and unsupervised image representation learning algorithms have achieved quantum leaps. However, these techniques do not account for representation resilience against noise in their design paradigms. Consequently, these effective methods suffer failure when confronted with noise outside the training distribution, such as complicated real-world noise that is usually opaque to model training. To address this issue, dual domains are optimized to separately model a canonical space for noisy representations, namely the Noise-Robust (NR) domain, and a twinned canonical clean space, namely the Noise-Free (NF) domain, by maximizing the interaction information between the representations. Given the dual canonical domains, we design a target-guided implicit neural mapping function to accurately translate the NR representations to the NF domain, yielding noise-resistant representations by eliminating noise regencies. The proposed method is a scalable module that can be readily integrated into existing learning systems to improve their robustness against noise. Comprehensive trials of various tasks using both synthetic and real-world noisy data demonstrate that the proposed Target-Guided Dual-Domain Translation (TDDT) method is able to achieve remarkable performance and robustness in the face of complex noisy images.

Recently, numerous end-to-end optimized image compression neural networks have been developed and proved themselves as leaders in rate-distortion performance. The main strength of these learnt compression methods is in powerful nonlinear analysis and synthesis transforms that can be facilitated by deep neural networks. However, out of operational expediency, most of these end-to-end methods adopt uniform scalar quantizers rather than vector quantizers, which are information-theoretically optimal. In this paper, we present a novel Lattice Vector Quantization scheme coupled with a spatially Adaptive Companding (LVQAC) mapping. LVQ can better exploit the inter-feature dependencies than scalar uniform quantization while being computationally almost as simple as the latter. Moreover, to improve the adaptability of LVQ to source statistics, we couple a spatially adaptive companding (AC) mapping with LVQ. The resulting LVQAC design can be easily embedded into any end-to-end optimized image compression system. Extensive experiments demonstrate that for any end-to-end CNN image compression models, replacing uniform quantizer by LVQAC achieves better rate-distortion performance without significantly increasing the model complexity., Comment: Accepted by CVPR 2023

Additional file 1: Table S1. Comparison of the TNM stages between Sarcopenia and No Sarcopenia groups in patients with gastric cancer. Table S2. Comparison of the TNM stages between Sarcopenia and No Sarcopenia groups in patients with colorectal cancer.

To enrich the functionalities of traditional cameras, light field cameras record both the intensity and direction of light rays, so that images can be rendered with user-defined camera parameters via computations. The added capability and flexibility are gained at the cost of gathering typically more than $100\times$ greater amount of information than conventional images. To cope with this issue, several light field compression schemes have been introduced. However, their ways of exploiting correlations of multidimensional light field data are complex and are hence not suited for inexpensive light field cameras. In this work, we propose a novel $\ell_\infty$-constrained light-field image compression system that has a very low-complexity DPCM encoder and a CNN-based deep decoder. Targeting high-fidelity reconstruction, the CNN decoder capitalizes on the $\ell_\infty$-constraint and light field properties to remove the compression artifacts and achieves significantly better performance than existing state-of-the-art $\ell_2$-based light field compression methods.

Lossless and near-lossless image compression is of paramount importance to professional users in many technical fields, such as medicine, remote sensing, precision engineering and scientific research. But despite rapidly growing research interests in learning-based image compression, no published method offers both lossless and near-lossless modes. In this paper, we propose a unified and powerful deep lossy plus residual (DLPR) coding framework for both lossless and near-lossless image compression. In the lossless mode, the DLPR coding system first performs lossy compression and then lossless coding of residuals. We solve the joint lossy and residual compression problem in the approach of VAEs, and add autoregressive context modeling of the residuals to enhance lossless compression performance. In the near-lossless mode, we quantize the original residuals to satisfy a given $\ell_\infty$ error bound, and propose a scalable near-lossless compression scheme that works for variable $\ell_\infty$ bounds instead of training multiple networks. To expedite the DLPR coding, we increase the degree of algorithm parallelization by a novel design of coding context, and accelerate the entropy coding with adaptive residual interval. Experimental results demonstrate that the DLPR coding system achieves both the state-of-the-art lossless and near-lossless image compression performance with competitive coding speed., Comment: arXiv admin note: text overlap with arXiv:2103.17015

Additional file 5: Table S5. Gene-specific primers used in RT-qPCR.

This thesis was scanned from the print manuscript for digital preservation and is copyright the author. Researchers can access this thesis by asking their local university, institution or public library to make a request on their behalf. Monash staff and postgraduate students can use the link in the References field.

Additional file 2: Table S2. Comparison of RNA-seq data with reference genome.

Additional file 3: Table S3. DEGs identified in vp5 and Vp5 embryos during maize seed maturation.

Additional file 1: Table S1. Statistical analysis of RNA-seq data quality.

Almost every single image restoration problem has a closely related parameter, such as the scale factor in super-resolution, the noise level in image denoising, and the quality factor in JPEG deblocking. Although recent studies on image restoration problems have achieved great success due to the development of deep neural networks, they handle the parameter involved in an unsophisticated way. Most previous researchers either treat problems with different parameter levels as independent tasks, and train a specific model for each parameter level; or simply ignore the parameter, and train a single model for all parameter levels. The two popular approaches have their own shortcomings. The former is inefficient in computing and the latter is ineffective in performance. In this work, we propose a novel system called functional neural network (FuncNet) to solve a parametric image restoration problem with a single model. Unlike a plain neural network, the smallest conceptual element of our FuncNet is no longer a floating-point variable, but a function of the parameter of the problem. This feature makes it both efficient and effective for a parametric problem. We apply FuncNet to super-resolution, image denoising, and JPEG deblocking. The experimental results show the superiority of our FuncNet on all three parametric image restoration tasks over the state of the arts., NeurIPS 2021

Based on a higher order finite-difference time-domain (FDTD(2,4)) method, and transmission line (TL) equations, an efficient time domain hybrid method is presented for the fast crosstalk analysis of multiple parallel microstrip lines. In this method, the per unit length (p.u.l) inductance and capacitance distribution parameters of multiple microstrip lines are solved by the empirical formulas firstly. Then, the crosstalk model of multiple parallel microstrip lines excited by lumped voltage source is established by the TL equations. Finally, the higher order FDTD (2,4) selecting large space step is utilized to discretize the TL equations to obtain the transient responses on these microstrip lines and terminal loads, which needs less mesh number and improves the computation efficiency without decreasing the calculation accuracy. A numerical example of three parallel microstrip lines excited by lumped voltage source is employed to verify the accuracy and efficiency of this presented method by comparing with the BLT equation and commercial software CST in terms of precision and time consumption. Moreover, the effects of length, height and thickness of microstrip lines on the crosstalk results are simulated and analyzed to master some valuable crosstalk principles.

Nighttime photographers are often troubled by light pollution of unwanted artificial lights. Artificial lights, after scattered by aerosols in the atmosphere, can inundate the starlight and degrade the quality of nighttime images, by reducing contrast and dynamic range and causing hazes. In this paper we develop a physically-based light pollution reduction (LPR) algorithm that can substantially alleviate the aforementioned degradations of perceptual quality and restore the pristine state of night sky. The key to the success of the proposed LPR algorithm is an inverse method to estimate the spatial radiance distribution and spectral signature of ground artificial lights. Extensive experiments are carried out to evaluate the efficacy and limitations of the LPR algorithm., 8 pages

A series of phase change materials (PCMs) microcapsules based on industrial paraffin core and calcium carbonate (CaCO3) shell were successfully synthesized for thermal energy storage. The SEM results revealed that the composites are mainly of spherical morphology. The FT-IR and XRD confirmed that the composites were mainly composed of the vaterite CaCO3, paraffin, and a small amount of calcite CaCO3. Compared with the composites prepared with sodium dodecyl sulfonate (SDS) as anionic surfactant, DSC and TG results of composites prepared with sodium dodecyl benzene sulfonate (SDBS) demonstrated that the encapsulation efficiency increased but the thermal stability decreased. Besides, with the different concentrations of CaCl2, the influence of anionic surfactants on the performance of microcapsules had been discussed. The results exhibited higher encapsulation efficiency of microcapsules with SDBS than that with SDS, even up to ten times. When the mass of CaCl2 is about 0.05 mol, encapsulation efficiency with SDBS can reach up to 20.80%. Accordingly, the super-cooling of PCM was significantly suppressed. From all results, it is worth developing the PCMs microcapsules with CaCO3 shell modulated by different anionic surfactants for thermal energy storage and SDBS is more suitable for preparing PCM@CaCO3 microcapsules.

A simple one-step interfacial polymerization method without special requirements for fabrication of paraffin@ polymethyl methacrylate (PCM@PMMA) nanocapsules is developed. Moreover, all the raw materials are cheap, innocuous and rich in resource. The influences of the mass ratio of methyl methacrylate/Paraffin (MMA/PCM) and stirring speed are studied. Scanning electron microscope (SEM), Fourier transformed infrared spectroscopy (FT-IR), Differential scanning calorimetric (DSC) and Thermo gravimetric analysis (TGA) are employed to characterize the as-prepared PCM@PMMA nanocapsules. Nano-encapsulated PCM could be successfully obtained under the stirring speed of 800 rpm for 2.5 h. The surface of the spherical nano-encapsulated PCMs is homogeneous, smooth and compact, and the particle size is in the range of 200–400 nm. The nano-encapsulated PCM are stable and reliable with the melting enthalpy of about 64.93 J/g and crystallization enthalpy of about 66.45 J/g. TGA results indicate that the nano-encapsulated PCM degraded in three steps and have good thermal stability with the content of the PCM about 52.95%. The PCM@PMMA nanocapsules could be promising for thermal energy storage application.

Objective. The current study aimed to explore the efficacy of Zero profile intervertebral fusion system (Zero-P) and traditional anterior plate cage system (PC) in the treatment of cervical spondylotic myelopathy (CSM). Further, the present study evaluated effects of the treatments on medical security, height of intervertebral disc, adjacent-level ossification development (ALOD), and adjacent segmentation disease (ASD) through a systematic retrospective analysis. Methods. Studies on Zero-P system and traditional anterior plate cage system for ACDF in the treatment of CSM were searched in PubMed, Web of Science, Ovid, Embase, and Cochrane Library databases. Two independent researchers screened articles, extracted data, and evaluated the quality of the articles based on the inclusion and exclusion criteria of the current study. RevMan5.3 software was used for meta-analysis following the guidelines of Cochrane collaboration network. Cervical curvature, interbody fusion rate, preoperative and postoperative disc height index (DHI), fusion cage sinking rate, postoperative dysphagia, ASD, ALOD, and loosening of screw were compared between the two groups. Results. A total of 17 literatures were included in the present study, including 6 randomized controlled trials and 11 observational studies. The studies comprised a total of 1204 patients with CSM, including 605 patients in the Zero-P system group (Zero-P group) and 599 patients in the traditional animal plate cage group (PC group). Results of this meta-analysis showed that postoperative dysphagia [OR = 0.40, CI (0.28, 95% 0.58), P

Validation of CRISPR-Cas9 editing typically explores the immediate vicinity of the gene editing site and distal off-target sequences, which has led to the conclusion that CRISPR-Cas9 editing is very specific. However, an increasing number of studies suggest that on-target unintended editing events like deletions and insertions are relatively frequent but unfortunately often missed in the validation of CRISPR-Cas9 editing. The deletions may be several kilobases-long and only affect one allele. The gold standard in molecular validation of gene editing is direct sequencing of relatively short PCR amplicons. This approach allows the detection of small editing events but fails in detecting large rearrangements, in particular when only one allele is affected. Detection of large rearrangements requires that an extended region is analyzed and the characterization of events may benefit from long-read sequencing. Here we implemented Xdrop™, a new microfluidic technology that allows targeted enrichment of long regions (~100 kb) using just a single standard PCR primer set. Sequencing of the enriched CRISPR-Cas9 gene-edited region in four cell lines on long- and short-read sequencing platforms unravelled unknown and unintended genome editing events. The analysis revealed accidental kilobases-large insertions in three of the cell lines, which remained undetected using standard procedures. We also applied the targeted enrichment approach to identify the integration site of a transgene in a mouse line. The results demonstrate the potential of this technology in gene editing validation as well as in more classic transgenics.

Background and Aims Stool DNA testing is an emerging and attractive option for colorectal cancer (CRC) screening. We previously evaluated the feasibility of a stool DNA (sDNA) test of methylated SDC2 for CRC detection. The aim of this study was to assess its performance in a multicenter clinical trial setting. Methods Each participant was required to undergo a sDNA test and a reference colonoscopy. The sDNA test consists of quantitative assessment of methylation status of SDC2 promoter. Results of real-time quantitative methylation-specific PCR were dichotomized as positive and negative, and the main evaluation indexes were sensitivity, specificity, and kappa value. All sDNA tests were performed and analyzed independently of colonoscopy. Results Among the 1110 participants from three clinical sites analyzed, 359 and 38 were diagnosed, respectively, with CRC and advanced adenomas by colonoscopy. The sensitivity of the sDNA test was 301/359 (83.8%) for CRC, 16/38 (42.1%) for advanced adenomas, and 134/154 (87.0%) for early stage CRC (stage I–II). Detection rate did not vary significantly according to age, tumor location, differentiation, and TNM stage, except for gender. The follow-up testing of 40 postoperative patients with CRC returned negative results as their tumors had been surgically removed. The specificity of the sDNA test was 699/713 (98.0%), and unrelated cancers and diseases did not seem to interfere with the testing. The kappa value was 0.84, implying an excellent diagnostic consistency between the sDNA test and colonoscopy. Conclusion Noninvasive sDNA test using methylated SDC2 as the exclusive biomarker is a clinically viable and accurate CRC detection method. Chinese Clinical Trial Registry Chi-CTR-TRC-1900026409, retrospectively registered on October 8, 2019; http://www.chictr.org.cn/edit.aspx?pid=43888&htm=4.

Validation of CRISPR-Cas9 editing typically explore the immediate vicinity of the gene editing site and distal off-target sequences, which have led to the conclusion that CRISPR-Cas9 editing is very specific. However, an increasing number of studies suggest that on-target unintended editing events like deletions and insertions are relatively frequent but unfortunately often missed in the validation of CRISPR-Cas9 editing. The deletions may be several kilobases-long and only affect one allele. The gold standard in molecular validation of gene editing is direct sequencing of relatively short PCR amplicons. This approach allows the detection of small editing events but fails in detecting large rearrangements, in particular when only one allele is affected. Detection of large rearrangements requires that an extended region is analyzed and the characterization of events may benefit from long-read sequencing. Here we implemented Xdrop™, a new microfluidic technology that allows targeted enrichment of long regions (~ 100 kb) using just a single standard PCR primer set. Sequencing of the enriched CRISPR-Cas9 gene edited region in 4 cell lines on long- and short -read sequencing platforms unravelled unknown and unintended genome editing events. The analysis revealed accidental kb large insertions in 3 of the cell lines, which remained undetected using standard procedures. We also applied the targeted enrichment approach to identify the integration site of a transgene in a mouse line. The results demonstrate the potential of this technology in gene editing validation as well as in more classic transgenics.

Whole slide imaging (WSI) is an emerging technology for digital pathology. The process of autofocusing is the main influence of the performance of WSI. Traditional autofocusing methods either are time-consuming due to repetitive mechanical motions, or require additional hardware and thus are not compatible to current WSI systems. In this paper, we propose the concept of \textit{virtual autofocusing}, which does not rely on mechanical adjustment to conduct refocusing but instead recovers in-focus images in an offline learning-based manner. With the initial focal position, we only perform two-shot imaging, in contrast traditional methods commonly need to conduct as many as 21 times image shooting in each tile scanning. Considering that the two captured out-of-focus images retain pieces of partial information about the underlying in-focus image, we propose a U-Net-inspired deep neural network based approach for fusing them into a recovered in-focus image. The proposed scheme is fast in tissue slides scanning, enabling a high-throughput generation of digital pathology images. Experimental results demonstrate that our scheme achieves satisfactory refocusing performance.

In many professional fields, such as medicine, remote sensing and sciences, users often demand image compression methods to be mathematically lossless. But lossless image coding has a rather low compression ratio (around 2:1 for natural images). The only known technique to achieve significant compression while meeting the stringent fidelity requirements is the methodology of $\ell_\infty$-constrained coding that was developed and standardized in nineties. We make a major progress in $\ell_\infty$-constrained image coding after two decades, by developing a novel CNN-based soft $\ell_\infty$-constrained decoding method. The new method repairs compression defects by using a restoration CNN called the $\ell_\infty\mbox{-SDNet}$ to map a conventionally decoded image to the latent image. A unique strength of the $\ell_\infty\mbox{-SDNet}$ is its ability to enforce a tight error bound on a per pixel basis. As such, no small distinctive structures of the original image can be dropped or distorted, even if they are statistical outliers that are otherwise sacrificed by mainstream CNN restoration methods. More importantly, this research ushers in a new image compression system of $\ell_\infty$-constrained encoding and deep soft decoding ($\ell_\infty\mbox{-ED}^2$). The $\ell_\infty \mbox{-ED}^2$ approach beats the best of existing lossy image compression methods (e.g., BPG, WebP, etc.) not only in $\ell_\infty$ but also in $\ell_2$ error metric and perceptual quality, for bit rates near the threshold of perceptually transparent reconstruction. Operationally, the new compression system is practical, with a low-complexity real-time encoder and a cascade decoder consisting of a fast initial decoder and an optional CNN soft decoder.

We present a CNN-based predictive lossless compression scheme for raw color mosaic images of digital cameras. This specialized application problem was previously understudied but it is now becoming increasingly important, because modern CNN methods for image restoration tasks (e.g., superresolution, low lighting enhancement, deblurring), must operate on original raw mosaic images to obtain the best possible results. The key innovation of this paper is a high-order nonlinear CNN predictor of spatial-spectral mosaic patterns. The deep learning prediction can model highly complex sample dependencies in spatial-spectral mosaic images more accurately and hence remove statistical redundancies more thoroughly than existing image predictors. Experiments show that the proposed CNN predictor achieves unprecedented lossless compression performance on camera raw images.

Single Image Super-Resolution (SISR) task refers to learn a mapping from low-resolution images to the corresponding high-resolution ones. This task is known to be extremely difficult since it is an ill-posed problem. Recently, Convolutional Neural Networks (CNNs) have achieved state of the art performance on SISR. However, the images produced by CNNs do not contain fine details of the images. Generative Adversarial Networks (GANs) aim to solve this issue and recover sharp details. Nevertheless, GANs are notoriously difficult to train. Besides that, they generate artifacts in the high-resolution images. In this paper, we have proposed a method in which CNNs try to align images in different spaces rather than only the pixel space. Such a space is designed using convex optimization techniques. CNNs are encouraged to learn high-frequency components of the images as well as low-frequency components. We have shown that the proposed method can recover fine details of the images and it is stable in the training process.

Additional file 1: Table S1. Summary of Bioinformatics Analysis of Example Datasets. Table S2. Primers Used for the Patient Samples. Figure S1. The most frequent integration site motif in the raw data of Cohn et al. The unfiltered raw dataset [1] comprising 80 million sites, was searched using MEME to find the most common short sequence motif, which is shown in sequence logo format aligned with the sequence of the LTR1, which is the first (outside) PCR primer used by [1] to amplify integration sites. Figure S2. Likely mispriming sites in the Cohn et al. integration dataset. Sequences (from hg19) that were within 50 bases of the 6719 integration sites determined by [1] (provided by the authors) were searched using BLAST for matches to their LTR1 primer. 1114 such sequences (± 30 bases of the reported integration site; arrow and dashed line) were found; 10 are shown aligned with LTR1 and with the matching bases highlighted in yellow. The chromosomal location of each site is shown to the left, with the number of patients reported to have a provirus at that site shown in parentheses. Figure S3. Plausible mechanisms for the erroneous identification of integration sites. The raw data used by [1] (NCBI accession number SRP045822) were searched for reads containing the cellular sequence 3′ of the reported integration site. No correct integration events were found, but several types of aberrant sequences were identified, including the 3 examples shown and labeled “raw read.” A. Double mispriming and PCR recombination. This sequence was most likely created by mispriming of LTR1 on the matching sequence (yellow) on chromosome 11 (magenta) as well as by LTR2 (green), and mispriming on a matching sequence on chromosome X (green), followed by recombination during PCR across the 8-base match indicated. B. Mispriming by a perfect fusion of sequences of LTR2-LTR1 on the same chromosome 11 sequence. C. Apparent correct integration (LTR2 followed by the 3′ 7 bases of the HIV-1 LTR1) two bases upstream of the reported integration site. The boxed sequence shows the 3′ most 7 nucleotides of the LTR, which are not in LTR2, but must be present in every correctly amplified integration site. Figure S4. The most common DNA sequence motif in the integration site datasets. Sequences ±50 base pairs of the reported integration sites from various studies are shown [1, 6, 11, 22], and 10,000 randomly chosen hg19 sequences and 10,000 Alu sequences were searched for common motifs using MEME [23]. The top hit in each case was aligned to the “integration site motif” of [1] (see text). The arrow above the sequence marks the site reported to be the preferred integration site [which is one nucleotide away from the site found at this site in the raw reads from patient 3, time point 3 (Figure S3). Figure S5. Sequence motifs adjacent to the integration site in various datasets. The patterns of preferred nucleotides in the host DNA immediately adjacent to the integration sites in patients reported by Maldarelli et al. [6] and Cohn et al. [1] were compared to data from cells infected in vitro (a PBMC dataset and a HeLa dataset). The sequence motifs were determined as previously described [8]. In the in vitro datasets, the target site nucleotides form a weak palindrome that matches what has been previously determined for HIV-1 using much smaller datasets [7, 8]. The preferred nucleotides in the Maldarelli dataset also match this motif. However, while the data of Cohn et al. shows some evidence of the palindrome; the sequence is weak and is not entirely symmetrical. The sequence from the Patient 3.3 sample (third time point) is obviously quite different from all the other data. Figure S6. Matching the Alu “integration sites” reported for the Patient 3.3 sample to the Alu consensus. The Alu consensus sequence is shown at the bottom. The quality of the match to the consensus is shown is the graph. The break in the match suggests the point in the sequence at which recombination frequently occurred.

In recent years, the chaos-based cryptographic algorithms have attracted a lot of attention. Specially, chaotic tent map (CTM)-based schemes show some good performances in randomness properties and security level. However, several shortcomings still can be found from them. In this paper, based on the security analysis of the pure CTM-based scheme, we propose a novel image encryption algorithm by using the combination of the rectangular transform and the CTM principle. It encrypts the three channels of the plain image at the same time, and these channel encryptions associate with each other. In addition, by generating the key streams related to both the secret keys and the plain image, its key-sensitivity has been further improved. The security of the proposed scheme has been verified by security analysis and experimental evaluations, and our results show that many drawbacks of pure CTM-based schemes have been overcome.

Given the success of the deep convolutional neural networks (DCNNs) in applications of visual recognition and classification, it would be tantalizing to test if DCNNs can also learn spatial concepts, such as straightness, convexity, left/right, front/back, relative size, aspect ratio, polygons, etc., from varied visual examples of these concepts that are simple and yet vital for spatial reasoning. Much to our dismay, extensive experiments of the type of cognitive psychology demonstrate that the data-driven deep learning (DL) cannot see through superficial variations in visual representations and grasp the spatial concept in abstraction. The root cause of failure turns out to be the learning methodology, not the computational model of the neural network itself. By incorporating task-specific convolutional kernels, we are able to construct DCNNs for spatial cognition tasks that can generalize to input images not drawn from the same distribution of the training set. This work raises a precaution that without manually-incorporated priors or features DCCNs may fail spatial cognitive tasks at rudimentary level.

Regularization techniques are widely used to improve the generality, robustness, and efficiency of deep convolutional neural networks (DCNNs). In this paper, we propose a novel approach of regulating DCNN convolutional kernels by a structured filter bank. Comparing with the existing regularization methods, such as $\ell_1$ or $\ell_2$ minimization of DCNN kernel weights and the kernel orthogonality, which ignore sample correlations within a kernel, the use of filter bank in regularization of DCNNs can mold the DCNN kernels to common spatial structures and features (e.g., edges or textures of various orientations and frequencies) of natural images. On the other hand, unlike directly making DCNN kernels fixed filters, the filter bank regularization still allows the freedom of optimizing DCNN weights via deep learning. This new DCNN design strategy aims to combine the best of two worlds: the inclusion of structural image priors of traditional filter banks to improve the robustness and generality of DCNN solutions and the capability of modern deep learning to model complex non-linear functions hidden in training data. Experimental results on object recognition tasks show that the proposed regularization approach guides DCNNs to faster convergence and better generalization than existing regularization methods of weight decay and kernel orthogonality., 11 pages, 19 figures

Lentiviral vectors (LV) have been used for the delivery of genes into hematopoietic stem and progenitor cells (HSPC) in clinical trials worldwide. LV, in contrast to retroviral vectors, have not been associated with insertion site-associated malignant clonal expansions, and thus have been considered safer. Here, however, we present a case of markedly abnormal dysplastic clonal hematopoiesis impacting the erythroid, myeloid and megakaryocytic lineages in a rhesus macaque transplanted with HSPCs that were transduced with a LV containing a strong retroviral murine stem cell virus (MSCV) constitutive promoter-enhancer in the LTR. 9 insertions were mapped in the abnormal clone, resulting in overexpression and aberrant splicing of several genes of interest, including the cytokine stem cell factor and the transcription factor PLAG1. This case represents the first clear link between a lentiviral insertion-induced clonal expansion and a clinically abnormal transformed phenotype following transduction of normal primate or human HSPC, and are thus concerning, and suggest that strong constitutive promoters should not be included within LV vectors.

Study purpose: Previously, we had a pilot study to evaluate the feasibility of a stool DNA test of methylated Syndycan-2 (SDC2) for colorectal cancer (CRC) detection. The purpose of this continuing investigation was to assess the performance of this test in a multicenter hospital-based setting. Methods: Each participant was required to undergo both a stool DNA test and a reference colonoscopy. The stool DNA test assesses the methylation status of a single target of SDC2 gene. Results of real-time methylation-specific PCR were dichotomized as positive and negative by a prespecified cut-off value. The main evaluation indexes were sensitivity, specificity, and kappa value, which was used to assess the diagnostic consistency between stool DNA test and colonoscopy. Tests were performed in blinded fashion and processed independently of the colonoscopic results. Results: Among the 1110 participants from three hospitals included for analysis, 359 and 38 had CRC and advanced adenomas, respectively, according to colonoscopy findings. The sensitivity of the stool DNA test was 83.8% for CRC and 42.1% for advanced adenomas. For CRC at early stage confined to bowel wall (I to II), the sensitivity was 87.0%. Detection rate for CRC did not vary significantly according to age, tumor location, differentiation, and TNM stage, except for gender (P=0.041). The specificity of the stool DNA test was 98.0% among participants with nonadvanced colorectal neoplasia, non-neoplastic disorders, or negative results on colonoscopy. Diagnostic consistency for CRC assessed by kappa value was 0.84, indicating that the agreement in outcome between the stool DNA test and colonoscopy was excellent. Conclusion: Noninvasive stool DNA test using methylated SDC2 as a molecular marker was a sensitive and accurate method for CRC detection. Citation Format: Jianping Wang, Side Liu, Hui Wang, Lei Zheng, Changchun Zhou, Guoxin Li, Rongkang Huang, Huaiming Wang, Chujun Li, Xinjuan Fan, Xinhui Fu, Xinying Wang, Hongliang Guo, Jie Guan, Yanlai Sun, Xilin Song, Zengjun Li, Dianbin Mu, Jujie Sun, Xianglin Liu, Yan Qi, Feng Niu, Chunhua Chen, Xiaolin Wu, Xianshu Wang, Xianrang Song, Hongzhi Zou. A multicenter clinical study of stool DNA test of methylated syndecan-2 for colorectal cancer detection [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4611.

Hepatocellular carcinoma (HCC) is an aggressive malignancy with its global incidence and mortality rate continuing to rise, although early detection and surveillance are suboptimal. We performed serological profiling of the viral infection history in 899 individuals from an NCI-UMD case-control study using a synthetic human virome, VirScan. We developed a viral exposure signature and validated the results in a longitudinal cohort with 173 at-risk patients who had long-term follow-up for HCC development. Our viral exposure signature significantly associated with HCC status among at-risk individuals in the validation cohort (area under the curve: 0.91 [95% CI 0.87-0.96] at baseline and 0.98 [95% CI 0.97-1] at diagnosis). The signature identified cancer patients prior to a clinical diagnosis and was superior to alpha-fetoprotein. In summary, we established a viral exposure signature that can predict HCC among at-risk patients prior to a clinical diagnosis, which may be useful in HCC surveillance.

Taking photos of optoelectronic displays is a direct and spontaneous way of transferring data and keeping records, which is widely practiced. However, due to the analog signal interference between the pixel grids of the display screen and camera sensor array, objectionable moir\'e (alias) patterns appear in captured screen images. As the moir\'e patterns are structured and highly variant, they are difficult to be completely removed without affecting the underneath latent image. In this paper, we propose an approach of deep convolutional neural network for demoir\'eing screen photos. The proposed DCNN consists of a coarse-scale network and a fine-scale network. In the coarse-scale network, the input image is first downsampled and then processed by stacked residual blocks to remove the moir\'e artifacts. After that, the fine-scale network upsamples the demoir\'ed low-resolution image back to the original resolution. Extensive experimental results have demonstrated that the proposed technique can efficiently remove the moir\'e patterns for camera acquired screen images; the new technique outperforms the existing ones.

All existing image enhancement methods, such as HDR tone mapping, cannot recover A/D quantization losses due to insufficient or excessive lighting, (underflow and overflow problems). The loss of image details due to A/D quantization is complete and it cannot be recovered by traditional image processing methods, but the modern data-driven machine learning approach offers a much needed cure to the problem. In this work we propose a novel approach to restore and enhance images acquired in low and uneven lighting. First, the ill illumination is algorithmically compensated by emulating the effects of artificial supplementary lighting. Then a DCNN trained using only synthetic data recovers the missing detail caused by quantization.

Recently a number of CNN-based techniques were proposed to remove image compression artifacts. As in other restoration applications, these techniques all learn a mapping from decompressed patches to the original counterparts under the ubiquitous $\ell_\infty$ metric. However, this approach is incapable of restoring distinctive image details which may be statistical outliers but have high semantic importance (e.g., tiny lesions in medical images). To overcome this weakness, we propose to incorporate an $\ell_\infty$ fidelity criterion in the design of neural network so that no small, distinctive structures of the original image can be dropped or distorted. Experimental results demonstrate that the proposed method outperforms the state-of-the-art methods in $\ell_\infty$ error metric and perceptual quality, while being competitive in $\ell_2$ error metric as well. It can restore subtle image details that are otherwise destroyed or missed by other algorithms. Our research suggests a new machine learning paradigm of ultra high fidelity image compression that is ideally suited for applications in medicine, space, and sciences., Accepted by DCC 2019

Taking photos of optoelectronic displays is a direct and spontaneous way of transferring data and keeping records, which is widely practiced. However, due to the analog signal interference between the pixel grids of the display screen and camera sensor array, objectionable moir�� (alias) patterns appear in captured screen images. As the moir�� patterns are structured and highly variant, they are difficult to be completely removed without affecting the underneath latent image. In this paper, we propose an approach of deep convolutional neural network for demoir��ing screen photos. The proposed DCNN consists of a coarse-scale network and a fine-scale network. In the coarse-scale network, the input image is first downsampled and then processed by stacked residual blocks to remove the moir�� artifacts. After that, the fine-scale network upsamples the demoir��d low-resolution image back to the original resolution. Extensive experimental results have demonstrated that the proposed technique can efficiently remove the moir�� patterns for camera acquired screen images; the new technique outperforms the existing ones.

Image of a scene captured through a piece of transparent and reflective material, such as glass, is often spoiled by a superimposed layer of reflection image. While separating the reflection from a familiar object in an image is mentally not difficult for humans, it is a challenging, ill-posed problem in computer vision. In this paper, we propose a novel deep convolutional encoder-decoder method to remove the objectionable reflection by learning a map between image pairs with and without reflection. For training the neural network, we model the physical formation of reflections in images and synthesize a large number of photo-realistic reflection-tainted images from reflection-free images collected online. Extensive experimental results show that, although the neural network learns only from synthetic data, the proposed method is effective on real-world images, and it significantly outperforms the other tested state-of-the-art techniques.

A novel composite PCMs nanocapsules (nanoPCMs) with PMMA and SiO2 as hybrid shell material and paraffin as core were successfully synthesized by concerted reaction of interfacial polymerization, sol–gel and self-assembly methods. SEM images showed that the structure of nanocapsules was affected by the mass ratio of TEOS and MMA. When the mass ratios of TEOS/MMA were 1:3.5 and 1:5.75, the nanocapsules were spherical and homogeneous with the diameter around 120 nm. However, the nanocapsules were inhomogeneous and irregular when the mass ratio of TEOS/MMA was 1:12.5. Based on the results of DSC, moderate mass ratio of TEOS/MMA can improve thermal properties of nanocapsules. When the mass ratio of TEOS/MMA was 1:3.5, the thermal properties of paraffin@PMMA–SiO2 nanocapsules was improved with the average enthalpy about 71 J/g and encapsulated efficiency of paraffin was 57.4%, which was higher than paraffin@PMMA nanocapsules. Besides paraffin which was encapsulated in nanocapsules has a better thermal stability than that of unencapsulated, and the nanocapsules prepared with different mass ratio of MMA/TEOS can provide almost the same thermal stability.

Despite great advances in lipid-lowering therapy, the gender-based disparities in lipid goal success among Chinese coronary heart disease (CHD) patients receiving lipid-lowering treatment remained uncharacterized. Furthermore, little is known about potential causes for this gender imbalance. This cross-sectional trial included 25 697 patients treated with lipid-lowering agents from 122 centres between April 2012 and October 2012; data for 9420 participants with CHD were analysed in the present study; all underwent clinical examination and had their latest fasting lipid profiles while on lipid-lowering therapy recorded. Logistic regression was used to explore possible factors associated with gender disparity in goal attainment rates, and to what extent each factor contributes. Of 9420 CHD patients (42.5% women) evaluated, women had a significantly lower overall low-density lipoprotein cholesterol (LDL-C) success rate than men (46.2 vs. 62.7%, P < 0.0001). Among CHD patients with diabetes, only 31.5% of women and 46.4% of men ( P < 0.0001) attained the optional LDL-C target of

Microbial-enhanced oil recovery (MEOR) is a biotechnology that utilizes biogas, biosurfactants, and biopolymers produced by, and degradation by subsurface fermentation process of microorganisms for enhancing oil production from substratum reservoir. In this report, we describe the progress made in pilot testing of MEOR in the Daqing oilfield of north China in recent decades and analyze suitable reservoir conditions and application requirements. By the end of 2012, cumulative oil increments reached 1.2 × 10 5 t using MEOR in the Daqing oilfield; of the oil produced, 518 wells of single-well microbial huff-and-puff yielded a cumulative incremental oil production of 6.3 × 10 4 t, and 10 projects with 45 well patterns adopted microbial flooding and profile modification to achieve a cumulative incremental oil production of 5.7 × 10 4 t. Thus, MEOR has proven to be an effective technique for routine operations in enhanced oil production in oil fields.

Background/Aims: Diabetic patients suffer from severe neointimal hyperplasia following angioplasty. The epigenetic abnormalities are increasingly considered to be relevant to the pathogenesis of diabetic cardiovascular complications. But the epigenetic mechanisms linking diabetes and coronary restenosis have not been fully elucidated. In this study, we explored the protective effect and underlying mechanisms of demethylases JMJD2A inhibition in balloon-injury induced neointimal formation in diabetic rats. Methods: JMJD2A inhibition was achieved by the chemical inhibitor 2,4-pyridinedicarboxylic acid (2,4-PDCA) and small interfering RNA (siRNA). In vitro, we investigated the proliferation, migration and inflammation of rat vascular smooth muscle cells (VSMCs) in response to high glucose (HG). In vivo, diabetic rats induced using high-fat diet and low-dose streptozotocin (35mg/kg) underwent carotid artery balloon injury. Morphometric analysis was performed using hematein eosin and immumohistochemical staining. Chromatin Immunoprecipitation (ChIP) was conducted to detect modification of H3K9me3 at inflammatory genes promoters. Results: The global JMJD2A was increased in HG-stimulated VSMCs and balloon-injured arteries of diabetic rats, accompanied by decreased H3K9me3. The inhibition of JMJD2A suppressed VSMCs proliferation, migration and inflammation induced by high glucose (HG) in vitro. And JMJDA2A inhibition attenuated neointimal formation in balloon-injured diabetic rats. The underlying mechanisms were relevant to the restoration of H3K9me3 levels at the promoters of MCP-1 and IL-6, and then the suppressed expression of MCP-1 and IL-6. Conclusion: The JMJD2A inhibition significantly attenuated neointimal formation in balloon injured diabetic rats via the suppression of VSMCs proliferation, migration, and inflammation by restoring H3K9me3.

The relationship between aerosol optical depth (AOD) and PM_(2.5) is often investigated in order to obtain surface PM_(2.5) from satellite observation of AOD with a broad area coverage. However, various factors could affect the AOD–PM_(2.5) regressions. Using both ground and satellite observations in Beijing from 2011 to 2015, this study analyzes the influential factors including the aerosol type, relative humidity (RH), planetary boundary layer height (PBLH), wind speed and direction, and the vertical structure of aerosol distribution. The ratio of PM_(2.5) to AOD, which is defined as η, and the square of their correlation coefficient (R^2) have been examined. It shows that η varies from 54.32 to 183.14, 87.32 to 104.79, 95.13 to 163.52, and 1.23 to 235.08 µg m^(−3) with aerosol type in spring, summer, fall, and winter, respectively. η is smaller for scattering-dominant aerosols than for absorbing-dominant aerosols, and smaller for coarse-mode aerosols than for fine-mode aerosols. Both RH and PBLH affect the η value significantly. The higher the RH, the smaller the η, and the higher the PBLH, the smaller the η. For AOD and PM2.5 data with the correction of RH and PBLH compared to those without, R^2 of monthly averaged PM_(2.5) and AOD at 14:00 LT increases from 0.63 to 0.76, and R^2 of multi-year averaged PM_(2.5) and AOD by time of day increases from 0.01 to 0.93, 0.24 to 0.84, 0.85 to 0.91, and 0.84 to 0.93 in four seasons respectively. Wind direction is a key factor for the transport and spatial–temporal distribution of aerosols originated from different sources with distinctive physicochemical characteristics. Similar to the variation in AOD and PM_(2.5), η also decreases with the increasing surface wind speed, indicating that the contribution of surface PM_(2.5) concentrations to AOD decreases with surface wind speed. The vertical structure of aerosol exhibits a remarkable change with seasons, with most particles concentrated within about 500 m in summer and within 150 m in winter. Compared to the AOD of the whole atmosphere, AOD below 500 m has a better correlation with PM_(2.5), for which R^2 is 0.77. This study suggests that all the above influential factors should be considered when we investigate the AOD–PM_(2.5) relationships.

A method for the facile preparation of oligoribonucleotide analogues containing β-d-allofuranosyl nucleosides with additional functional groups tethered to the 6′-O positions is presented. It is based on the synthesis of two protected nucleosides carrying a 6′-O-bromopentyl and a 6′-O-methylaminopentyl substituent. By a simple two-step procedure, these key intermediates were transformed into two phosphoramidites carrying a 1-aza-18-crown-6 and a triethyleneglycol group, respectively, each capable of complexing metal ions. By automated synthesis, these functionalized nucleoside analogues were efficiently incorporated into short oligoribonucleotides. Under physiological conditions (150 mM NaCl, 2 mM MgCl2, pH 7.4), incorporation of a single allofuranosyl cytosine substituted with a triethyleneglycol moiety led to a significant enthalpic stabilization of an A-type RNA duplex. This observation is in agreement with a metal ion-mediated stabilizing interaction between the two pairing strands

This paper presents a generic pre-processor for expediting conventional template matching techniques. Instead of locating the best matched patch in the reference image to a query template via exhaustive search, the proposed algorithm rules out regions with no possible matches with minimum computational efforts. While working on simple patch features, such as mean, variance and gradient, the fast pre-screening is highly discriminative. Its computational efficiency is gained by using a novel octagonal-star-shaped template and the inclusion-exclusion principle to extract and compare patch features. Moreover, it can handle arbitrary rotation and scaling of reference images effectively. Extensive experiments demonstrate that the proposed algorithm greatly reduces the search space while never missing the best match.

Small compression noises, despite being transparent to human eyes, can adversely affect the results of many image restoration processes, if left unaccounted for. Especially, compression noises are highly detrimental to inverse operators of high-boosting (sharpening) nature, such as deblurring and superresolution against a convolution kernel. By incorporating the non-linear DCT quantization mechanism into the formulation for image restoration, we propose a new sparsity-based convex programming approach for joint compression noise removal and image restoration. Experimental results demonstrate significant performance gains of the new approach over existing image restoration methods.

In November 2016 we submitted to arXiv our paper "Automated Inference on Criminality Using Face Images". It generated a great deal of discussions in the Internet and some media outlets. Our work is only intended for pure academic discussions; how it has become a media consumption is a total surprise to us. Although in agreement with our critics on the need and importance of policing AI research for the general good of the society, we are deeply baffled by the ways some of them mispresented our work, in particular the motive and objective of our research.

This article is a sequel to our earlier work [25]. The main objective of our research is to explore the potential of supervised machine learning in face-induced social computing and cognition, riding on the momentum of much heralded successes of face processing, analysis and recognition on the tasks of biometric-based identification. We present a case study of automated statistical inference on sociopsychological perceptions of female faces controlled for race, attractiveness, age and nationality. Our empirical evidences point to the possibility of training machine learning algorithms, using example face images characterized by internet users, to predict perceptions of personality traits and demeanors.

In order to develop the peripheral part of Daqing oilfield, one mixed bacteria system was screened and used for microbial flooding on Microbial Enhanced Oil Recovery (MEOR) process in Daqing low permeability oilfield where there are huge resource potentials but many disadvantageous conditions. From 2004 to 2009, Microbial flooding field test was conducted in Chaoyanggou oilfield. In 2004, Microbial flooding pilot test with 2 injectors and 10 producers was carried out. Seven out of 10 oil wells showed effective results, with 14.5 thousand tons of accumulative incremental oil. In October 2009, expanding field test with 9 injectors and 24 producers was carried out in the same block. By now, 15 oil wells have shown good effect. The daily oil production increased from 32.4 tons to 43.2 tons and the incremental oil amounts to 9370 tons. The microbial flooding results showed that wells along the main channel sand and developed fracture zones had a better response.

Using Fourier transform near infrared (FT-NIR) spectroscopy, the characteristics and phase change mechanism of the phase change process for PMMA-SiO2@PCM microcapsules were explored. We investigated changes in the microstructure of the microcapsules as well. The results show that the melting of paraffin molecules in the microcapsules is a process wherein the-CH2 symmetric stretching vibration gradually increases and the asymmetric stretching vibration changes randomly. During the paraffin phase change process, the change in absorbance intensity is only half that of the shell materials. Additionally, NIR spectroscopy was used to analyze the core-shell structure and to monitor the phase change process for the microcapsules. The application of NIR spectroscopy to study the phase change process for microcapsule phase change materials has scientific significance and is of value to study the phase change mechanism for the determination of the most efficient phase change materials.