Visual Sentiment Analysis with Network in Network

Zuhe Li,Yangyu Fan,Fengqin Wang 보안공학연구지원센터 2016 International Journal of Signal Processing, Image Vol.9 No.9

In modern society, visual content like images and videos is increasingly becoming a new form of media to express users’ opinions on the Internet. As a complement to textual sentiment analysis, visual sentiment analysis intends to provide more robust information for data analytics by extracting emotion and sentiment toward topics and events from images and videos. Inspired by recent works that applied deep convolutional neural networks (CNN) to this challenging problem, we proposed a framework for image sentiment analysis with a novel deep neural network called Network in Network (NIN) which intends to improve the discriminability for local patches within receptive fields. We trained our network on a dataset consisting of nearly half a million Flickr images and minimized the effect of noisy training data by fine-tuning the network in a progressive manner. Extensive experiments conducted on manually labeled Twitter images show that the proposed architecture performs better in visual sentiment analysis than conventional CNN and other traditional algorithms.

Unsupervised Feature Learning Assisted Visual Sentiment Analysis

Zuhe Li,Yangyu Fan,Fengqin Wang,Weihua Liu 보안공학연구지원센터 2016 International Journal of Multimedia and Ubiquitous Vol.11 No.10

Visual sentiment analysis which aims to understand the emotion and sentiment in visual content has attracted more and more attention. In this paper, we propose a hybrid approach for visual sentiment concept classification with an unsupervised feature learning architecture called convolutional autoencoder. We first extract a representative set of unlabeled patches from the image dataset and discover useful features of these patches with sparse autoencoders. Then we use a convolutional neural network (CNN) to obtain feature activations on full images for sentiment concept classification. We also fine-tune the network with a progressive strategy in order to filter out noisy samples in the weakly labeled training data. Meanwhile, we use low-level visual features to classify visual sentiment concepts in a traditional manner. At last the classification results with unsupervised feature learning and that with traditional features are taken into account together with a fusion algorithm to make a final prediction. Extensive experiments on benchmark datasets reveal that the proposed approach can achieve better performance in visual sentiment analysis compared to its predecessors.

Mathematical Modelling of Online Warping Height of Cold-Rolled Thin Strip Steel

Yanglong Li,Jie Wen,Haihai Lin,Meng Yu,Fengqin Wang 한국강구조학회 2022 International Journal of Steel Structures Vol.22 No.4

The warping deformation is a common shape defect for thin strip steel. The warping form is shown as type-C for online strip with tension, whereas it is generally type-L for offl ine steel sheet without tension. A mathematical model between type- C warping height for online strip and type-L warping height for offl ine steel sheet was established and validated by fi eld experiments. The predicted offl ine warping height was agreed well with the measured results, and the average relative error was about 5.4%. There is a linear relationship between online and offl ine warping heights. The infl uence factors on online warping height were also discussed. At the same offl ine warping value of steel sheet, the online warping value increases with the increase of warping correction factor and the decrease of the sampling sheet length. The online warping height is also aff ected by width and Poisson’s ratio of strip. This model could be used to predict the online warping height of steel strip based on the offl ine warping height of steel sheet. Also the offl ine warping height can be evaluated according to the online warping state. It lays a primary theoretical foundation for the warping relationship between online strip and offl ine steel sheet.

AMPK alleviates high uric acid-induced Na+-K+-ATPase signaling impairment and cell injury in renal tubules

Jing Xiao,Sibo Zhu,Haochen Guan,Haochen Guan,Fengqin Li,Xiaoli Zhang,Hui Guo,Xiaojun Wang,Zhibin Ye 생화학분자생물학회 2019 Experimental and molecular medicine Vol.51 No.-

One of the mechanisms in hyperuricemia (HUA)-induced renal tubular injury is the impairment of Na+-K+-ATPase (NKA) signaling, which further triggers inflammation, autophagy, and mitochondrial dysfunction and leads to cell injury. Here, we used RNA sequencing to screen the most likely regulators of NKA signaling and found that the liver kinase B1(LKB1)/adenosine monophosphate (AMP)-activated protein kinase (AMPK)/ mammalian target of rapamycin (mTOR) pathway was the most abundantly enriched pathway in HUA. AMPK is a key regulator of cell energy metabolism; hence, we examined the effect of AMPK on HUA-induced dysregulation of NKA signaling and cell injury. We first detected AMPK activation in high uric acid (UA)-stimulated proximal tubular epithelial cells (PTECs). We further found that sustained treatment with the AMPK activator 5-aminoimidazole-4-carboxamide 1-β-d-ribofuranoside (AICAR), but not the AMPK inhibitor Compound C, significantly alleviated UA-induced reductions in NKA activity and NKA α1 subunit expression on the cell membrane by reducing NKA degradation in lysosomes; sustained AICAR treatment also significantly alleviated activation of the NKA downstream molecules Src and interleukin-1β (IL-1β) in PTECs. AICAR further alleviated high UA-induced apoptosis, autophagy, and mitochondrial dysfunction. Although AMPK activation by metformin did not reduce serum UA levels in hyperuricemic rats, it significantly alleviated HUAinduced renal tubular injury and NKA signaling impairment in vivo with effects similar to those of febuxostat. Our study suggests that AMPK activation may temporarily compensate for HUA-induced renal injury. Sustained AMPK activation could reduce lysosomal NKA degradation and maintain NKA function, thus alleviating NKA downstream inflammation and protecting tubular cells from high UA-induced renal tubular injury.

Integrative applications of network pharmacology and molecular docking: An herbal formula ameliorates H9c2 cells injury through pyroptosis

Zhongwen Qi,Zhipeng Yan,Yueyao Wang,Nan Ji,Xiaoya Yang,Ao Zhang,Meng Li,Fengqin Xu,Junping Zhang The Korean Society of Ginseng 2023 Journal of Ginseng Research Vol.47 No.2

Background: QiShen YiQi pills (QSYQ) is a Traditional Chinese Medicine (TCM) formula, which has a significant effect on the treatment of patients with myocardial infarction (MI) in clinical practice. However, the molecular mechanism of QSYQ regulation pyroptosis after MI is still not fully known. Hence, this study was designed to reveal the mechanism of the active ingredient in QSYQ. Methods: Integrated approach of network pharmacology and molecular docking, were conducted to screen active components and corresponding common target genes of QSYQ in intervening pyroptosis after MI. Subsequently, STRING and Cytoscape were applied to construct a PPI network, and obtain candidate active compounds. Molecular docking was performed to verify the binding ability of candidate components to pyroptosis proteins and oxygen-glucose deprivation (OGD) induced cardiomyocytes injuries were applied to explore the protective effect and mechanism of the candidate drug. Results: Two drug-likeness compounds were preliminarily selected, and the binding capacity between Ginsenoside Rh2 (Rh2) and key target High Mobility Group Box 1 (HMGB1)was validated in the form of hydrogen bonding. 2 μM Rh2 prevented OGD-induced H9c2 death and reduced IL-18 and IL-1β levels, possibly by decreasing the activation of the NLRP3 inflammasome, inhibiting the expression of p12-caspase1, and attenuating the level of pyroptosis executive protein GSDMD-N. Conclusions: We propose that Rh2 of QSYQ can protect myocardial cells partially by ameliorating pyroptosis, which seems to have a new insight regarding the therapeutic potential for MI.