Search

Most facial expression recognition (FER) algorithms are based on shallow features, and the deep networks tend to lose some key features in the expression, such as eyes, nose and mouth. To address the limitations, we present in this paper a novel approach, named CBAM-Global-Efficient Channel Attention-ResNet (C-G-ECA-R). C-G-ECA-R combines a strong attention mechanism and residual network. The strong attention enhances the extraction of important features of expressions by embedding the channel and spatial attention mechanism before and after the residual module. The addition of Global-Efficient Channel Attention (G-ECA) into the residual module strengthens the extraction of key features and reduces the loss of facial information. The extensive experiments have been conducted on two publicly available datasets, Extended Cohn-Kanade and Japanese Female Facial Expression. The results demonstrate that our proposed C-G-ECA-R, especially under ResNet34, has achieved 98.98% and 97.65% accuracy, respectively for the two datasets, that are higher than the state-of-arts. [ABSTRACT FROM AUTHOR]

The Xinbaishiyan tunnel in the reconstruction Chengdu–Kunming railway Ermeishan-Mipan section mainly runs through dolomite with dolomitic limestone, with an excavation area of 260 m2, a maximum span of 22.3 m, a maximum height of 14.4 m, a vector height of 7 m, and a rise-span ratio of 0.31. The tunnel has an extra-large cross section, and it is a low flat-ration railway tunnel. This paper mainly describes the finite element analysis for this tunnel excavation that was used to guide the construction. Finite element software was used to model the tunnel according to the engineering geological conditions of the tunnel. These engineering geological conditions included the rock mass, system bolts, middle pipe shed, steel arch and shotcrete, grouting layer, second lining and so on. Nonlinear construction phase analysis was adopted. The results showed that the maximum vertical deformation of the tunnel vault and the middle of the invert was about 34 mm. The vertical deformation of the tunnel could be divided into an acceleration deformation section, linear deformation section, deceleration deformation section, and stable deformation section. The maximum horizontal deformation in the middle of the side wall was about 12.3 mm. Under the effect of the initial support, the equivalent stress of the side wall gradually increased with the excavation of the steps and the increase in the support structure. The axial force of the bolt in the middle of the side wall was larger than that in other places and the axial force of the middle pipe shed went along with the excavation of the tunnel in waves. The steel arch and the shotcrete had the maximum effective stress at the arch shoulder, which played the role of the deformation and pressure for the surrounding rock. During the construction, the length and height of the three-step method had to be set reasonably. The middle pipe shed and the system bolt supported the rock mass together. In the construction of the extra-large cross section and the flat tunnel, there was no need to set up temporary support, which was convenient for mechanical excavation. [ABSTRACT FROM AUTHOR]

In order to study the influence of different spacing and angles of the axes of the two tunnels on the deformation and operation safety of existing tunnel in near-tunnel construction, the project in Fuzhou is taken to analysis. We use the finite element analysis software to construct a three-dimensional finite element model and carry on the numerical simulation, obtaining vertical vault displacement curves under different spacing and angles of the axes of the two tunnels. Then, we compare it with the actual monitoring data and do further analysis, we get following conclusions: (1) the main performance of the existing tunnel deformation is upward, the deformation decreases with the increase of the spacing and crossing angle, and the effect of the spacing on the existing tunnel vault deformation is far greater than the crossing angle. (2) When the spacing is less than 1.5 times of the new tunnel hole diameter, the maximum tunnel vault deformation occurs in the tunnel cross surface; when the spacing is greater than 1.5 times of the new tunnel hole diameter, the maximum tunnel vault deformation gradually moves toward to the center of the existing tunnel. The research results can provide some references for the design and construction of similar projects. [ABSTRACT FROM AUTHOR]

Wear and thermal cracking are the major reasons for the fatigue of a hot-work die; remarkable improvement has been achieved by laser biomimetic treatment in recent years. However, because of severe service conditions, the wear resistance and thermal fatigue behavior of a hot-work die should be further improved. Therefore, pulse current stimulation (PCS) was used to treat H13 steel processed by laser biomimetic treatment. The results indicate that the microstructures of the laser-melted zone and matrix were refined by PCS, thus increasing the hardness. Therefore, the wear resistance of laser biomimetic sample (LBS) was improved by PCS. The tensile strength of LBS at room temperature and 650°C increased with PCS. The residual compressive stress distributed on the surface of LBS is induced by PCS. Therefore, the LBS treated by PCS has the best thermal fatigue behavior. [ABSTRACT FROM AUTHOR]

Thrust and torque have been selected in this research for online detection and measurements of drill wear for the drilling of stainless steel parts because cutting forces are closely related to the drilling process and give the best indirect indication of drill conditions. Using thrust and torque information to monitor the drill conditions for control of the drilling process can decrease the operation cost and enhance the product quality. To find the most important feature(s), the feature selection technique is used in this research. Sequential forward search algorithm is used for feature selection. To reduce the dimension of the measurement vector, it is necessary to retain only those components of the extracted features which show a high sensitivity to drill wear and low sensitivity to process parameters. The best feature selected is the peak of torque in the drilling process. Adaptive neuro-fuzzy inference system (ANFIS) is a neuro-fuzzy system. It includes input layer, output layer, and layers between them. ANFIS can construct fuzzy rules with membership functions to generate an input–output pair. A 1 × 6 ANFIS architecture with generalized bell membership function can achieve a success rate of 100 % for the online detection of drill states. A 1 × 6 ANFIS architecture with product of sigmoid membership functions can measure the drill wear online with an error as low as 0.15 %. Furthermore, the detection and measurement of drill wear is performed under different drilling conditions as compared with the training process. This shows that ANFIS has the capability of generalization. [ABSTRACT FROM AUTHOR]