Search

Abstract: This paper presents a neural model which learns low-dimensional nonlinear manifolds embedded in higher-dimensional data space based on mixtures of local linear manifolds under a self-organizing framework. Compared to other similar networks, the local linear manifolds learned by our network have a more localized representation of local data distributions thanks to a new distortion measure, which removes confusion between sub-models that exists in many similar mixture models. Each neuron in the network asymptotically learns a mean vector and a principal subspace of the data in its local region. It is proved that there is no local extremum for each sub-model. Experiments show that the new mixture model is better adapted to nonlinear manifolds of various data distributions than other similar models. The online-learning property of this model is desirable when the data set is very large, when computational efficiency is of paramount importance, or when data are sequentially input. We further show an application of this model to recognition of handwritten digit images based on mixtures of local linear manifolds. [Copyright &y& Elsevier]

Single image dehazing is a challenging ill-posed problem which estimates latent haze-free images from observed hazy images. Some existing deep learning based methods are devoted to improving the model performance via increasing the depth or width of convolution. The learning ability of Convolutional Neural Network (CNN) structure is still under-explored. In this paper, a Detail-Enhanced Attention Block (DEAB) consisting of Detail-Enhanced Convolution (DEConv) and Content-Guided Attention (CGA) is proposed to boost the feature learning for improving the dehazing performance. Specifically, the DEConv contains difference convolutions which can integrate prior information to complement the vanilla one and enhance the representation capacity. Then by using the re-parameterization technique, DEConv is equivalently converted into a vanilla convolution to reduce parameters and computational cost. By assigning the unique Spatial Importance Map (SIM) to every channel, CGA can attend more useful information encoded in features. In addition, a CGA-based mixup fusion scheme is presented to effectively fuse the features and aid the gradient flow. By combining above mentioned components, we propose our Detail-Enhanced Attention Network (DEA-Net) for recovering high-quality haze-free images. Extensive experimental results demonstrate the effectiveness of our DEA-Net, outperforming the state-of-the-art (SOTA) methods by boosting the PSNR index over 41 dB with only 3.653 M parameters. (The source code of our DEA-Net is available at https://github.com/cecret3350/DEA-Net.).

The purpose of image dehazing is to remove the interference from weather factors in degraded images and enhance the clarity and color saturation of images to maximize the restoration of useful features. Single image dehazing is one of the most important tasks in the field of image restoration. In recent years, due to the progress of deep learning, single image dehazing has made great progress. With the success of Transformer in advanced computer vision tasks, some research studies also began to apply Transformer to image dehazing tasks and obtained surprising results. However, both the deconvolution-neural-network-based dehazing algorithm and Transformer based dehazing algorithm magnify their advantages and disadvantages separately. Therefore, this paper proposes a novel Transformer-Convolution fusion dehazing network (TCFDN), which uses Transformer's global modeling ability and convolutional neural network's local modeling ability to improve the dehazing ability. In the Transformer-Convolution fusion dehazing network, the classic self-encoder structure is used. This paper proposes a Transformer-Convolution hybrid layer, which uses an adaptive fusion strategy to make full use of the Swin-Transformer and convolutional neural network to extract and reconstruct image features. On the basis of previous research, this layer further improves the ability of the network to remove haze. A series of contrast experiments and ablation experiments not only proved that the Transformer-Convolution fusion dehazing network proposed in this paper exceeded the more advanced dehazing algorithm, but also provided solid and powerful evidence for the basic theory on which it depends.

Inventory is the basis of business activities; inventory management helps industries keep their inventories stocked with reasonable quantities, which ensures consumers demand while minimizing storage costs. The traditional manual inventory management has low efficiency and a high labor cost. In this paper, we used improved YOLOv3 to detect the cups stored on the warehouse shelves and counted their numbers to realize automated inventory management. The warehouse images are collected by the camera and transmitted to the industrial computer, which runs the YOLOv3 network. There are three feature maps in YOLOv3, the two smaller feature maps and the structure behind them are removed, and the k-means algorithm is used to optimize the default anchor size. Moreover, the detection range is limited to a specified area. Experiments show that, by eliminating those two feature maps, the network parameter is reduced from 235 MB to 212 MB, and detection FPS is improved from 48.15 to 54.88 while mAP is improved from 95.65% to 96.65% on our test dataset. The new anchors obtained by the k-means algorithm further improve the mAP to 96.82%. With those improvements, the average error rate of detection is reduced to 1.61%. Restricted detection areas eliminate irrelevant items to ensure the high accuracy of the detection result. The accurately counted number of cups and its change provide significant data for inventory management.

Esophageal cancer is an exceedingly aggressive and malignant cancer that imposes a substantial burden on patients and their families. It is usually treated with surgery, chemotherapy, radiotherapy, and molecular-targeted therapy. Immunotherapy is a novel treatment modality for esophageal cancer wherein genetically engineered adoptive cell therapy is utilized, which modifies immune cells to attack cancer cells. Using chimeric antigen receptor (CAR) or T cell receptor (TCR) modified T cells yielded demonstrably encouraging efficacy in patients. CAR-T cell therapy has shown robust clinical results for malignant hematological diseases, particularly in B cell-derived malignancies. Natural killer (NK) cells could serve as another reliable and safe CAR engineering platform, and CAR-NK cell therapy could be a more generalized approach for cancer immunotherapy because NK cells are histocompatibility-independent. TCR-T cells can detect a broad range of targeted antigens within subcellular compartments and hold great potential for use in cancer therapy. Numerous studies have been conducted to evaluate the efficacy and feasibility of CAR and TCR based adoptive cell therapies (ACT). A comprehensive understanding of genetically-modified T cell technologies can facilitate the clinical translation of these adoptive cell-based immunotherapies. Here, we systematically review the state-of-the-art knowledge on genetically-modified T-cell therapy and provide a summary of preclinical and clinical trials of CAR and TCR-transgenic ACT., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2021 Zhu, Xiao, Zhang, Cui, Lu and Wu.)

Chimeric antigen receptor T (CAR-T) cell therapy has exhibited a substantial clinical response in hematological malignancies, including B-cell leukemia, lymphoma, and multiple myeloma. Therefore, the feasibility of using CAR-T cells to treat solid tumors is actively evaluated. Currently, multiple basic research projects and clinical trials are being conducted to treat lung cancer with CAR-T cell therapy. Although numerous advances in CAR-T cell therapy have been made in hematological tumors, the technology still entails considerable challenges in treating lung cancer, such as on-target, of-tumor toxicity, paucity of tumor-specific antigen targets, T cell exhaustion in the tumor microenvironment, and low infiltration level of immune cells into solid tumor niches, which are even more complicated than their application in hematological tumors. Thus, progress in the scientific understanding of tumor immunology and improvements in the manufacture of cell products are advancing the clinical translation of these important cellular immunotherapies. This review focused on the latest research progress of CAR-T cell therapy in lung cancer treatment and for the first time, demonstrated the underlying challenges and future engineering strategies for the clinical application of CAR-T cell therapy against lung cancer., Competing Interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest., (Copyright © 2021 Xiao, Zhang, Zhu, Cui, Lu, Yu and Wu.)

Aim: To understand the implication of GATA-4 and GATA-5 methylation in gastric carcinogenesis., Methods: Methylation status of GATA-4 and GATA-5 CpG islands in human gastric mucosa samples, including normal gastric biopsies from 45 outpatients, gastric dysplasia [low-grade gastric intraepithelial neoplasia (GIN), n = 30; indefinite, n = 77], and 80 paired sporadic gastric carcinomas (SGC) as well as the adjacent non-neoplastic gastric tissues was analyzed by methylation specific polymerase chain reaction (MSP) and confirmed by denatured high performance liquid chromatography (DHPLC). Immunohistochemical staining was used to detect protein expression. The correlation between GATA-4 and GATA-5 methylation and clinicopathological characteristics of patients including Helicobacter pylori (H. pylori) infection was analyzed., Results: GATA-4 and GATA-5 methylation was frequently observed in SGCs (53.8% and 61.3%, respectively) and their corresponding normal tissues (41.3% and 46.3%) by MSP. The result of MSP was consistent with that of DHPLC. Loss of both GATA-4 and GATA-5 proteins was associated with their methylation in SGCs (P = 0.01). Moreover, a high frequency of GATA-4 and GATA-5 methylation was found in both gastric low-grade GIN (57.1% and 69.0%) and indefinite for dysplasia (42.9% and 46.7%), respectively. However, GATA-4 and GATA-5 methylation was detected only in 4/32 (12.5%) and 3/39 (7.7%) of normal gastric biopsies. GATA-4 methylation in both normal gastric mucosa and low-grade GIN was also significantly associated with H. pylori infection (P = 0.023 and 0.027, two-sides)., Conclusion: Epigenetic inactivation of GATA-4 (and GATA-5) by methylation of CpG islands is an early frequent event during gastric carcinogenesis and is significantly correlated with H. pylori infection.

Background: Suicide gene therapy is a widely used molecular treatment for head and neck cancer. In this study, we try to use the method of homogenous recombination in bacteria to clone thymidine kinase gene (tk)-a kind of suicide gene to adenovirus backbone vectors for the construction of replication-defective adenoviruses., Methods: pAdTrack-CMV/tk was constructed through subclone of a restriction endonuclease fragment including thymidine kinase gene from plasmid pCMV-tk to another plasmid pAdTrack-CMV, and then co-transfected with supercoiled pAdEasy-1, which was an adenoviral backbone vector except for deletions of E1 and E3, to competent E. coli BJ5183 for homogenous recombination using electroporation procedure. With the same method, pAdTrack-CMV was also co-transformed with pAdEasy-1 for homogenous recombination in BJ5183. Identified with restriction endonuclease PacI and polymerase chain reaction (PCR), plasmids pAd-GFP/tk and pAd-GFP were successfully constructed. Each of them was digested with PacI and sequently transfected into human embryo kidney 293 cells (HEK293) using Lipofectamine 2000., Results: Comet-like adenovirus-producing foci of Ad-GFP/tk and Ad-GFP were observed after 5 to 7 days of cell culture. After twelve days of packaging, the replication-defective adenoviruses were collected. Identified with PCR, thymidine kinase gene was successfully constructed into Ad-GFP/tk., Conclusion: The replication-defective adenoviruses containing thymidine kinase can be constructed more easily by homogenous recombination in bacteria than conventional techniques.

The rapid growth of digital multimedia and Internet technologies has made copyright protection, copy protection, and integrity verification three important issues in the digital world. To solve these problems, the digital watermarking technique has been presented and widely researched. Traditional watermarking algorithms are mostly based on discrete transform domains, such as the discrete cosine transform, discrete Fourier transform (DFT), and discrete wavelet transform (DWT). Most of these algorithms are good for only one purpose. Recently, some multipurpose digital watermarking methods have been presented, which can achieve the goal of content authentication and copyright protection simultaneously. However, they are based on DWT or DFT. Lately, several robust watermarking schemes based on vector quantization (VQ) have been presented, but they can only be used for copyright protection. In this paper, we present a novel multipurpose digital image watermarking method based on the multistage vector quantizer structure, which can be applied to image authentication and copyright protection. In the proposed method, the semi-fragile watermark and the robust watermark are embedded in different VQ stages using different techniques, and both of them can be extracted without the original image. Simulation results demonstrate the effectiveness of our algorithm in terms of robustness and fragility.

Objective: To investigate the association between the single nucleotide polymorphisms (SNPs) of interferon regulatory factor 3 (IRF-3) gene and the susceptibility of esophageal cancer in Anyang area, Henan Province., Methods: The genomic DNAs of 152 esophageal cancer patients and 191 health controls were extracted from the peripheral blood leukocytes. Three fragments covering codons 96, 194, 377, and 427 of the IRF-3 gene were amplified by polymerase chain reaction (PCR). The SNPs were revealed by DNA sequencing and the genotype of every sample was determined accordingly. The frequency distribution of the SNPs was analyzed by chi(2) test., Results: There were no IRF-3 gene SNPs at codons 96, 194, or 377 in Anyang participants. There was significant difference in the SNPs at codon 427 between the healthy controls and esophageal cancer patients in Anyang area OR=2.38 (95%CI=1.15-4.95,P<0.05)., Conclusion: The SNPs at codon 427 of the IRF-3 gene may relate to the susceptibility of esophageal cancer. The risk of esophageal cancer in participants with C allele is 2.38-fold compared to those with G allele.

Objective: To investigate the relationship between human papillomavirus (HPV) infection and the development of upper digestive tract carcinomas., Methods: In situ hybridization (ISH) technique was used to detect the expression of HPV16 E6 mRNA in 40 specimens of esophageal squamous cell carcinoma, 74 specimens of gastric cardiac adenocarcinoma, 40 specimens of adenocarcinoma of gastric corpus, 62 specimens of adenocarcinoma of gastric antrum, all resected during operation, and 58 specimens of normal tissues of gastric mucosa, obtained by endoscopy., Results: The positive rate of HPV16 E6 mRNA was 70% (28/40) in esophageal squamous cell carcinomas, 67.6% (50/74) in gastric cardiac adenocarcinomas, 47.5% (19/40) in adenocarcinoma of gastric corpus, 35.5% (22/62) in adenocarcinoma of gastric antrum, and 20% (10/50) in normal tissues of gastric mucosa respectively, without a significant difference between esophageal squamous cell carcinoma and gastric cardiac adenocarcinoma (P = 0.955). The positive rate in gastric adenocarcinoma was significantly lower than that in gastric cardiac adenocarcinoma (P < 0.05) or in esophageal squamous cell carcinomas (P < 0.01). The positive rate in normal tissue of gastric mucosa was significantly lower than those in gastric adenocarcinoma (P < 0.05) in gastric cardiac adenocarcinoma (P < 0.01), and in esophageal squamous cell carcinoma (P < 0.01). The positive rate in gastric cardiac carcinomas was negatively correlated with TNM stage (P < 0.05). However, there was no correlation between HPV16 and gender, age, tumor size, depth of penetration, differentiation, lymph node metastasis, vascular invasion and prognosis of these patients with gastric cardiac carcinoma (all P > 0.05)., Conclusion: HPV infects not only squamous epithelium, but also glandular epithelium. HPV16 may be an etiologic factor in the development of upper digestive tract carcinomas, especially esophageal and cardiac carcinomas.

Aim: To investigate the putative role of human papillomavirus (HPV) infection in the carcinogenesis of esophageal squamous cell carcinoma in China., Methods: Twenty-three esophageal squamous cell carcinoma samples and the distal normal epithelium from Shanxi Province, and 25 more esophageal squamous cell carcinoma samples from Anyang city, two areas with a high incidence of esophageal cancer in China, were detected for the existence of HPV-16 DNA by PCR, mRNA in situ hybridization (ISH) and immunohistochemistry (IHC) targeting HPV-16 E6 gene., Results: There were approximately 64 % (31/48) patients having HPV-16 DNA in tumor samples, among them nearly two-thirds (19/31) samples were detected with mRNA expression of HPV-16 E6. However, in the normal esophageal epithelium from cancer patients, the DNA and mRNA of HPV-16 were found with much less rate: 34.7 % (8/23) and 26.1 % (6/23) respectively. In addition, at protein level detected by IHC assay, 27.1 % (13/48) tumor samples had virus oncoprotein E6 expression, while only one case of normal epithelium was found positive., Conclusion: HPV infection, especially type 16, should be considered as a risk factor for esophageal malignancies in China.

In this paper, a new and fast encoding algorithm for vector quantization is presented. This algorithm makes full use of two characteristics of a vector: the sum and the variance. A vector is separated into two subvectors: one is composed of the first half of vector components and the other consists of the remaining vector components. Three inequalities based on the sums and variances of a vector and its two subvectors components are introduced to reject those codewords that are impossible to be the nearest codeword, thereby saving a great deal of computational time, while introducing no extra distortion compared to the conventional full search algorithm. The simulation results show that the proposed algorithm is faster than the equal-average nearest neighbor search (ENNS), the improved ENNS, the equal-average equal-variance nearest neighbor search (EENNS) and the improved EENNS algorithms. Comparing with the improved EENNS algorithm, the proposed algorithm reduces the computational time and the number of distortion calculations by 2.4% to 6% and 20.5% to 26.8%, respectively. The average improvements of the computational time and the number of distortion calculations are 4% and 24.6% for the codebook sizes of 128 to 1024, respectively.