• Journal of Information Processing Systems
• /
• v.20 no.4
• /
• pp.535-549
• /
• 2024

Facial expressions (FEs) serve as fundamental components for human emotion assessment and human-computer interaction. Traditional convolutional neural networks tend to overlook valuable information during the FE feature extraction, resulting in suboptimal recognition rates. To address this problem, we propose a deep learning framework that incorporates hierarchical feature fusion, contextual data, and an attention mechanism for precise FE recognition. In our approach, we leveraged an enhanced VGGNet16 as the backbone network and introduced an improved group convolutional channel attention (GCCA) module in each block to emphasize the crucial expression features. A partial decoder was added at the end of the backbone network to facilitate the fusion of multilevel features for a comprehensive feature map. A reverse attention mechanism guides the model to refine details layer-by-layer while introducing contextual information and extracting richer expression features. To enhance feature distinguishability, we employed islanding loss in combination with softmax loss, creating a joint loss function. Using two open datasets, our experimental results demonstrated the effectiveness of our framework. Our framework achieved an average accuracy rate of 74.08% on the FER2013 dataset and 98.66% on the CK+ dataset, outperforming advanced methods in both recognition accuracy and stability.

• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.556-570
• /
• 2021

The existing video expression recognition methods mainly focus on the spatial feature extraction of video expression images, but tend to ignore the dynamic features of video sequences. To solve this problem, a multi-mode convolution neural network method is proposed to effectively improve the performance of facial expression recognition in video. Firstly, OpenFace 2.0 is used to detect face images in video, and two deep convolution neural networks are used to extract spatiotemporal expression features. Furthermore, spatial convolution neural network is used to extract the spatial information features of each static expression image, and the dynamic information feature is extracted from the optical flow information of multiple expression images based on temporal convolution neural network. Then, the spatiotemporal features learned by the two deep convolution neural networks are fused by multiplication. Finally, the fused features are input into support vector machine to realize the facial expression classification. Experimental results show that the recognition accuracy of the proposed method can reach 64.57% and 60.89%, respectively on RML and Baum-ls datasets. It is better than that of other contrast methods.

• Journal of Microbiology and Biotechnology
• /
• v.16 no.7
• /
• pp.1026-1031
• /
• 2006

The optimization of the production of a fusion protein containing the antigenic domain 1 (AD-1) is of a great importance, considering its use in diagnostic tests. The fusion protein is produced by the fermentation of a recombinant strain of Escherichia coli containing the plasmid Mbg58, which expresses the AD-1 (aa 484-650) of human cytomegalovirus glycoprotein B as a fusion protein together with aa 1-375 of ${\beta}-galactosidase$. An important characteristic of promoters (lac and derivatives) used in recombinant protein production in E. coli is their inducibility. Induction by IPTG is widely used for basic research; however, its use in large-scale production is undesirable because of its high cost and toxicity. In this work, studies using different inducers and carbon sources for the production of a fusion protein containing the AD-l were performed. The results showed that lactose could be used as an inducer in the fermentation process for the production of this protein, and that expression levels could exceed those achieved with IPTG. The use of lactose for protein expression in E. coli should be extremely useful for the inexpensive, large-scale production of heterologous proteins in E. coli. Addition of sucrose to the fermentation medium improved the yield of recombinant protein, whereas addition of fructose or trehalose decreased the yield.

• KSBB Journal
• /
• v.16 no.1
• /
• pp.1-10
• /
• 2001

Escherichia coli has been used as an expression work horse for foreign genes. This article summarized recent development in genetic engineering techniques for overproduction of medical proteins and industrial enzymes. Special emphasis was placed upon research activities concerning folding and refolding of inclusion bodies at genetic and fermentation levels. Plasmid and mRNA stabilization, development of strong inducible promoters, modification of translational elements and reduction of rpoteolytic degradation were carried out to elevate an expression level of a target protein. Optimization of culture conditions, improvement of denaturation and renaturation steps and coexpression of molecular chaperones or foldase were accomplished to produce active proteins in soluble form. Fusion protein systems with selective separation and surface display technology were also performed in an effort to make the E. coli expression system more effective and versatile.

• Polymer(Korea)
• /
• v.35 no.4
• /
• pp.289-295
• /
• 2011

Biomaterials for retinal tissue engineering must demonstrate several critical features for potential utility, including mechanical integrity, biocompatibility, and slow biodegradation. Silk film biomaterials were designed and characterized to meet these functional requirements. We prepared natural/synthetic hybrid silk/PLGA films using 0, 10, 20, 40, and 80 wt% of silk by a solvent evaporation method. MIT assay was used to confirm the number of cells attached on film at 1, 2, and 3 days, respectively. The morphology of cellular adhesion on films was also confirmed by scanning electron microscope (SEM). RT-PCR was conducted to confrrm mRNA expression of retinal pigment epithelitun (RPE) using RPE65 as a RPEs marker and the expression of cytokeratin were determined by immunofluorescence staining. We confirmed that the silk/PLGA film of 20~40 wt% silk was superior for the adhesion and proliferation of RPEs.

• Molecules and Cells
• /
• v.42 no.3
• /
• pp.262-269
• /
• 2019

The porcine myeloid antimicrobial peptide (PMAP), one of the cathelicidin family members, contains small cationic peptides with amphipathic properties. We used a putative lysozyme originated from the bacteriophage P22 (P22 lysozyme) as a fusion partner, which was connected to the N-terminus of the PMAP36 peptide, to markedly increase the expression levels of recombinant PMAP36. The PMAP36-P22 lysozyme fusion protein with high solubility was produced in Escherichia coli. The final purified yield was approximately 1.8 mg/L. The purified PMAP36-P22 lysozyme fusion protein exhibited antimicrobial activity against both Gram-negative and Grampositive bacteria (Staphylococcus aureus, Salmonella enterica serovar Typhimurium, Pseudomonas aeruginosa, and Bacillus subtilis). Furthermore, we estimated its hemolytic activity against pig erythrocytes as 6% at the high concentration ($128{\mu}M$) of the PMAP36-P22 lysozyme fusion protein. Compared with the PMAP36 peptide (12%), our fusion protein exhibited half of the hemolytic activity. Overall, our recombinant PMAP36-P22 lysozyme fusion protein sustained the antimicrobial activity with the lower hemolytic activity associated with the synthetic PMAP36 peptide. This study suggests that the PMAP36-P22 lysozyme fusion system could be a crucial addition to the plethora of novel antimicrobials.

• Korean Chemical Engineering Research
• /
• v.47 no.5
• /
• pp.624-629
• /
• 2009

It is the most important step to screen soluble and insoluble proteins when we attempt to improve the solubility of recombinant proteins through directed evolution approach. Here we show that the solubility of a recombinant protein in vivo can be examined by expressing the recombinant protein with beta-lactamase as a fusion partner. First we constructed an expression system which can produc a fusion protein with the C-terminal of beta-lactamase. Two soluble proteins, i.e. adenine deaminase and aspartate aminotransferase, and insoluble GlcNAc-2-epimerase were cloned into the developed expression vector, respectively. We investigated the effect of the expression of the three recombinant fusion proteins on the growth of E. coli, and confirmed that the solubilities of the recombinant proteins correlated with cell growth rates.

• Biomedical Science Letters
• /
• v.2 no.2
• /
• pp.231-240
• /
• 1996

Human blood clotting (coagulation) factor 9 cDNA which codes for 461 amino acid has been cloned by screening human fetal liver cDNA library using PCR. This 1.4 kb cDNA spanning from the ATG initiation codon to the TAA termination codon was cloned into bacterial .expression vector pGEX-2T, generating pGEX-F9 plasmid. The plasmid pGEX-F9 expresses about 73 kDa GST (Glutathione S-transferase)-Factor 9 fusion protein when introduced into E. coli. Western blot analysis using polyclonal antibody raised against human factor 9 confirmed this fusion protein contains factor 9 protein. The level of GST-factor 9 expression was about 20% of total protein and the purification of fusion protein was efficiently achieved by using GST agarose bead based on one step purification protocol.

• Korean Chemical Engineering Research
• /
• v.56 no.5
• /
• pp.711-717
• /
• 2018

Human epidermal growth factor (hEGF) can stimulate the division of various cell types and has potential clinical applications. Since the protein contains three intra-molecular disulfide bonds, the high expression of active hEGF in Escherichia coli has not been well researched, We fused the hEGF gene with a small ubiquitin-related modifier gene (SUMO) by synthesizing an artificial SUMO-hEGF fusion gene that was highly expressed in E. coli (DE3) strain. The optimal expression level of the soluble fusion protein, SUMO-hEGF with IPTG (Isopropyl-${\beta}$-D-Thiogalactopyranoside), was up to 38.9% of the total cellular protein. The fusion protein was purified by Ni-NTA affinity chromatography and cleaved by a SUMO-specific protease to obtain the native hEGF, which was further purified by Ni-NTA affinity chromatography. The result of the reverse-phase HPLC showed that the purity of the recombinant cleaved hEGF was greater than 98%.

• Journal of Microbiology and Biotechnology
• /
• v.25 no.2
• /
• pp.274-279
• /
• 2015

Receptor activator of nuclear factor-kappa B ligand (RANKL) is a critical factor in osteoclastogenesis. It makes osteoclasts differentiate and multinucleate in bone remodeling. In the present study, RANKL was expressed as a soluble maltose binding protein (MBP)-fusion protein using the Escherichia coli maltose binding domain tag system (pMAL) expression vector system. The host cell E. coli DH5α was cultured and induced by isopropyl β-_D-1-thiogalactopyranoside for rRANKL expression. Cells were disrupted by sonication to collect soluble MBP-fused rRANKL. The MBP-fusion rRANKL was purified with MBP Trap affinity chromatography and treated with Tobacco Etch Virus nuclear inclusion endopeptidase (TEV protease) to remove the MBP fusion protein. Dialysis was then carried out to remove binding maltose from the cleaved rRANKL solution. The cleaved rRANKL was purified with a second MBP Trap affinity chromatography to separate unsevered MBP-fusion rRANKL and cleaved MBP fusion protein. The purified rRANKL was shown to have biological activity by performing in vitro cell tests. In conclusion, biologically active rRANKL was successfully purified by a simple two-step chromatography purification process with one column.