• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권9호
• /
• pp.2505-2528
• /
• 2023

The field of facial expression recognition (FER) has been actively researched to improve human-computer interaction. In recent years, deep learning techniques have gained popularity for addressing FER, with numerous studies proposing end-to-end frameworks that stack or widen significant convolutional neural network layers. While this has led to improved performance, it has also resulted in larger model sizes and longer inference times. To overcome this challenge, our work introduces a novel lightweight model architecture. The architecture incorporates three key factors: Depth-wise Separable Convolution, Residual Block, and Attention Modules. By doing so, we aim to strike a balance between model size, inference speed, and accuracy in FER tasks. Through extensive experimentation on popular benchmark FER datasets, our proposed method has demonstrated promising results. Notably, it stands out due to its substantial reduction in parameter count and faster inference time, while maintaining accuracy levels comparable to other lightweight models discussed in the existing literature.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권13호
• /
• pp.5271-5276
• /
• 2014

Mitogen-activated protein kinase/extracellular signal-regulated kinase kinase kinase 3 (MEKK3) is an important protein kinase and a member of the MAPK family, which regulates cellular responses to environmental stress and serves as key integration points along the signal transduction cascade that not only link diverse extracellular stimuli to subsequent signaling molecules but also amplify the initiating signals to ultimately activate effector molecules and induce cell proliferation, differentiation and survival. To explore the relationship between MEKK3 and cell apoptosis, clinicopathology and prognosis, we characterize the expression of MEKK3 and survivin in cervical cancer. MEKK3 and survivin expression was measured by RT-PCR and Western blotting of fresh surgical resections from 30 cases of cervical cancer and 25 cases of chronic cervicitis. Protein expression was detected by tissue microarray and immunochemistry (En Vision) in 107 cases of cervical cancer, 86 cases of cervical intraepithelial neoplasia (CIN), and 35 cases of chronic cervicitis. Expression patterns were analyzed for their association with clinicopathological factors and prognosis in cervical cancer. Expression of MEKK3 and survivin mRNA was significantly higher in cervical cancer than in the controls (p<0.05). MEKK3 and survivin expression differed significantly between cervical carcinoma, CIN, and cervicitis (p<0.05) and correlated with clinical stage, infiltration depth, and lymph node metastasis (p<0.05). MEKK3 expression was positively correlated with survivin (p<0.05). Kaplan-Meier survival analysis showed that MEKK3 and survivin expression, lymph node metastasis, depth of invasion, and FIGO stage reduce cumulative survival. Cox multivariate regression analysis showed that MEKK3, survivin, and clinical staging are independent prognostic factors in cervical cancer (p<0.05). Expression of MEKK3 and survivin are significantly increased in cervical cancer, their overexpression participating in the occurrence and development of cervical cancer, with protein expression and clinical staging acting as independent prognostic factors for patients with cervical cancer.

• Asian Pacific Journal of Cancer Prevention
• /
• 제16권15호
• /
• pp.6789-6794
• /
• 2015

Background: To analyze the expression of estrogen receptors (ER), progesterone receptors (PR), C-erbB-2 and Ki-67 in endometrial carcinoma (EC) and their relationships with the clinicopathological features. Materials and Methods: Sixty-seven EC samples, 53 normal endometrial samples and 53 atypical hyperplasia endometrial samples were all selected in Shaanxi Provincial People's Hospital from Jun., 2012 to Jun., 2014. The expression of ER, PR, C-erbB-2 and Ki-67 in EC tissue, normal endometrial tissue and atypical hyperplasia endometrial tissue was respectively detected using immunohistochemical SP method. The relationships between the expression of ER, PR, C-erbB-2 and Ki-67 and the patients' clinicopathological features as well as their correlations in EC tissue were also analyzed. Results: The positive expression rates of ER and PR in EC tissue were 44.8% and 41.8%, respectively, dramatically lower than in atypical hyperplasia endometrial tissue and normal endometrial tissue (P<0.01). The positive expression rates of C-erbB-2 and Ki-67 in EC tissue were 80.6% and 64.2%, respectively, significantly higher than in atypical hyperplasia endometrial tissue and normal endometrial tissue (P<0.01). In EC tissue, the expression of ER and PR was closely associated with the differentiated degrees and depth of myometrial invasion (P<0.05), while that of C-erbB-2 and Ki-67 with the clinical staging, differentiated degrees, depth of myometrial invasion and presence or absence of lymph node metastasis (P<0.05). Spearman correlation analysis further displayed that the expression of ER was positively correlated with PR (r=0.393, P=0.001), but negatively with C-erbB-2 and Ki-67 (r=-0.469, P=0.000; r=-0.329, P=0.007); The expression of PR was negatively correlated with C-erbB-2 and Ki-67 (r=-0.273, P=0.025; r=-0.251, P=0.041), but that of C-erbB-2 positively with Ki-67 (r=0.342, P=0.005). Conclusions: Abnormal expression of ER, PR, C-erbB2 and Ki-67 might play an important role in endometrial malignant transformation and cell differentiation, so their joint detection is likely to be a comprehensive combination of immune factors, which is of great importance for EC prognosis.

• 원예과학기술지
• /
• 제35권2호
• /
• pp.243-251
• /
• 2017

Plants have evolved to adapt to abiotic stresses, such as salt, cold, and drought stress. In this study, we conducted an in-depth analysis of drought resistance mechanisms by constructing a gene co-expression network in Chinese cabbage (Brassica rapa ssp. pekinensis L.). This drought stress co-expression network has 1,560 nodes, 4,731 edges, and 79 connected components. Based on genes that showed significant co-expression in the network, drought tolerance was associated with the induction of reactive oxygen species removal by raffinose family oligosaccharides and inositol metabolism. This network could be a useful tool for predicting the functions of genes involved in drought stress resistance in Chinese cabbage.

• Asian Pacific Journal of Cancer Prevention
• /
• 제16권7호
• /
• pp.2845-2849
• /
• 2015

Overexpression of interleukin (IL)-35 has been found in a variety of malignancies, but the expression status in gastric cancer has yet to be elucidated clearly. In the present study, positive expression of EBI3 and p35 was 63.3% and 70.0% of cases, respectively. EBI3 expression was strongly related with larger tumor size and invasion depth (P<0.05). Similarly, expression of p35 was also correlated with larger tumor size (P<0.05). These results indicate that IL-35 might be involved in growth of gastric cancer. Interestingly, EBI3 and p35 expressions were positive correlated with Ki-67 expression. Moreover, EBI3 immunoreactivity was associated with Bcl-2 staining. Our data suggest IL-35 is correlated with genesis of gastric cancer by regulating growth and apoptosis.

• Journal of Periodontal and Implant Science
• /
• 제52권5호
• /
• pp.383-393
• /
• 2022

Purpose: Aloe-emodin (AE), a natural anthraquinone abundant in aloe plants and rhubarb (Rheum rhabarbarum), has long been used to treat chronic inflammatory diseases. However, AE's underlying mechanisms in periodontal inflammation have not been fully elucidated. Acidic mammalian chitinase (AMCase) is a potential biomarker involved in bone remodeling. This study aimed to evaluate AE's effect on periodontitis in rats and investigate AMCase expression. Methods: Eighteen Sprague-Dawley rats were separated into the following groups: healthy (group 1), disease (group 2), vehicle (group 3), AE high-dose (group 4), and AE low-dose (group 5). Porphyromonas gingivalis ligatures were placed in rats (groups 2-5) for 7 days. Groups 4 and 5 were then treated with AE for an additional 14 days. Saliva was collected from all groups, and probing pocket depth was measured in succession. Periodontal pocket tissues were subjected to histomorphometric analysis after the rats were sacrificed. Bone marrow-derived macrophages and murine macrophages were stimulated with receptor activator of nuclear factor-κB ligand (RANKL) and treated with different concentrations of AE. AMCase expression was detected from the analysis of saliva, periodontal pocket tissues, and differentiated osteoclasts. Results: Among rats with P. gingivalis-induced periodontitis, the alveolar bone resorption levels and periodontal pocket depth were significantly reduced after treatment with AE. AMCase protein expression was significantly higher in the disease group than in the healthy control (P<0.05). However, AE inhibited periodontal inflammation by downregulating AMCase expression in saliva and periodontal pocket tissue. AE significantly reduced RANKL-stimulated osteoclastogenesis by modulating AMCase (P<0.05). Conclusions: AE decreases alveolar bone loss and periodontal inflammation, suggesting that this natural anthraquinone has potential value as a novel therapeutic agent against periodontal disease.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권22호
• /
• pp.9661-9666
• /
• 2014

Background: A number of studies have identified a shared susceptibility locus in phospholipase C epsilon 1 (PLCE1) for esophageal squamous cell carcinoma (ESCC) and gastric cardia adenocarcinomas (GCA). However, the results of PLCE1 expression in esophageal and gastric cancer remain inconsistent and controversial. Moreover, the effects on clinicopathological features remain undetermined. This study aimed to provide a precise quantification of the association between PLCE1 expression and the risk of ESCC and GCA through meta-analysis. Materials and Methods: Eligible studies were identified from PubMed, Wanfang Data, ISI Web of Science, and the Chinese National Knowledge Infrastructure databases. Using RevMan5.2 software, pooled odds ratios (ORs) with 95% confidence intervals (CIs) were employed to assess the association of PLCE1 expression with clinicopathological features relative to ESCC or GCA. Results: Seven articles were identified, including 761 esophageal and gastric cancer cases and 457 controls. Overall, we determined that PLCE1 expression was associated with tumor progression in both esophageal cancers (pooled OR=5.93; 95%CI=3.86 to 9.11) and gastric cancers (pooled OR=9.73; 95%CI=6.46 to 14.7). Moreover, invasion depth (pooled OR=3.62; 95%CI=2.30 to 5.70) and lymph node metastasis (pooled OR=4.21; 95%CI=2.69 to 6.59) were linked with PLCE1 expression in gastric cancer. However, no significant associations were determined between PLCE1 overexpression and the histologic grade, invasion depth, and lymph node metastasis in esophageal cancer. Conclusions: Our metaanalysis results indicated that upregulated PLCE1 is significantly associated with an increased risk of tumor progression in ESCC and GCA. Therefore, PLCE1 expression can be appropriately regarded as a promising biomarker for ESCC and GCA patients.

• Asian Pacific Journal of Cancer Prevention
• /
• 제16권18호
• /
• pp.8653-8658
• /
• 2016

Background: Heparanase is believed to be involved in gastric carcinogenesis. However, the clinicopathologic features of gastric cancer with high heparanase expression remain unclear. Aim : The purpose of this study was to comprehensively and quantitatively summarize available evidence for the use of heparanase mRNA and protein expression to evaluate the clinicopathological associations in gastric cancer in Asian patients by meta-analysis. Materials and Methods: Relevant articles listed in MEDLINE, CNKI and the Cochrane Library databases up to MARCH 2015 were searched by use of several keywords in electronic databases. A meta-analysis was performed to clarify the impact of heparanase mRNA and protein on clinicopathological parameters in gastric cancer. Combined ORs with 95%CIs were calculated by Revman 5.0, and publication bias testing was performed by stata12.0. Results: A total of 27 studies which included 3,891 gastric cancer patients were combined in the final analysis. When stratifying the studies by the pathological variables of heparanase mRNA expression, the depth of invasion (633 patients) (OR=4.96; 95% CI=2.38-1.37; P<0.0001), lymph node metastasis (639 patients) (OR=6.22; 95%CI=2.70-14.34, P<0.0001), and lymph node metastasis (383 patients) (OR=6.85; 95% CI=2.04-23.04; P=0.002) were all significant. When stratifying the studies by the pathological variables of heparanase protein expression, this was the case for depth of invasion (1250 patients) (OR=2.76; 95% CI=1.52-5.03; P=0.0009), lymph node metastasis (1178 patients) (OR=4.79 ; 95% CI=3.37-6.80, P<0.00001), tumor size (727 patients) (OR=2.06 ; 95% CI=1.31-3.23; P=0.002) (OR=2.61; 95% CI=2.09-3.27; P=0.000), and TNM stage (1233 patients) (OR=6.85; 95% CI=2.04-23.04; P=0.002). Egger's tests suggested publication bias for depth of invasion, lymph node metastasis, lymph node metastasis and tumor size of heparanase mRNA and protein expression. Conclusions: This meta-analysis suggests that higher heparanase expression in gastric cancer is associated with clinicopathologic features of depth of invasion, lymph node metastasis and TNM stage at mRNA and protein levels, and of tumor size only at the protein level. Egger's tests suggested publication bias for these clinicopathologic features of heparanase mRNA and protein expression, and which may be caused by shortage of relevant studies. As a result, although abundant reports showed heparanase may be associated with clinicopathologic features in gastric cancer, this meta-analysis indicates that more strict studies were needed to evaluate its clinicopathologic significance.

• Asian Pacific Journal of Cancer Prevention
• /
• 제15권2호
• /
• pp.989-997
• /
• 2014

AMFR, autocrine motility factor receptor, also called gp78, is a cell surface cytokine receptor which has a dual role as an E3 ubiquitin ligase in endoplasmic reticulum-associated degradation. AMFR expression is associated with tumor malignancy. We here investigated the clinical significance of AMFR and its role in metastasis and prognosis in gastric cancer. Expression of AMFR, E-cadherin and N-cadherin in cancer tissues and matched adjacent normal tissues from 122 gastric cancer (GC) patients undergoing surgical resection was assessed by immunohistochemistry. Levels of these molecules in 17 cases selected randomly were also analysed by Western blotting. AMFR expression was significantly increased in gastric cancer tissues, and associated with invasion depth and lymph node metastasis. Kaplan-Meier analysis showed AMFR expression correlated with poor overall survival and an increased risk of recurrence in the GC cases. Cox regression analysis suggested AMFR to be an independent predictor for overall and recurrence-free survival. E-cadherin expression was decreased in gastric cancer tissues; conversely, N-cadherin was increased. Expression of AMFR negatively correlated with E-cadherin expression, whereas N-cadherin expression showed a significant positive correlation with AMFR expression. AMFR might be involved in the regulation of epithelial-mesenchymal transition, with aberrant expression correlating with a poor prognosis and promoting invasion and metastasis in GCs.

• 전기학회논문지
• /
• 제65권10호
• /
• pp.1731-1737
• /
• 2016

In this paper, a method is proposed to recognize human actions as nonverbal expression; the proposed method is composed of two steps which are action representation and action recognition. First, MHI(Motion History Image) is used in the action representation step. This method includes segmentation based on depth information and generates spatio-temporal templates to describe actions. Second, CNN(Convolution Neural Network) which includes feature extraction and classification is employed in the action recognition step. It extracts convolution feature vectors and then uses a classifier to recognize actions. The recognition performance of the proposed method is demonstrated by comparing other action recognition methods in experimental results.