Search

Urban fringe areas, characterized by relatively larger community sizes and lower population densities compared to central areas, may lead to variations in walkability as well as gender differences, such as safety perception. While objective measurements have received considerable attention, further research is needed to comprehensively assess subjective perceptions of walking in the urban periphery. As a case study, we evaluated survey responses of community perceptions of "Imageability", "Enclosure", "Human scale", "Complexity" and "Safety" of Shanghai's five new towns, comparing these with responses from the central area in terms of gender difference, and analyzed influencing factors and prediction performance of machine learning (ML) models. We developed a TrueSkill-based rating system to dynamically collect audits of street view images (SVIs) from professional students and used the result to integrate with Geographic Information Systems (GIS), Computer Vision (CV), Clustering analysis, and ML algorithm for further investigation. Results show that most of the new towns' communities are perceived as moderately walkable or higher, with the city center's community exhibiting the best walkability perceptions in general. Male and female perceptions of the "Human scale" and the factors that affect it differ little, but there are significant disparities in the other four perceptions. The best-performing ML models were effective at variable explanations and generalizations, with Random Forest Regression (RFR) performing better on more perception predictions. Responses also suggest that certain street design factors, such as street openness, can positively influence walkability perceptions of women and could be prioritized in new town development and urban renewal for more inclusive and walkable cities. [ABSTRACT FROM AUTHOR]

To solve the problem of low classification accuracy caused by differences in object types, shapes, and scales in SAS images, an object classification method based on a deformable residual network and transfer learning is proposed. First, a lightweight deformable convolution module DSDCN was designed by adding offsets to a traditional convolution, to adapt to objects with different shapes in SAS images, and the depthwise separable convolution was used to optimize the module. Second, a deformable residual network was designed with the DSDCN, which combined the traditional depth features with deformable features for object representation and improved the robustness of the model. Furthermore, the network was trained by the transfer learning method to save training time and prevent model overfitting. The model was trained and validated on the acquired SAS images. Compared with other existing state-of-the art models, the classification accuracy in this study improved by an average of 6.83% and had an advantage in the amount of computation, which is 108 M. On the deformation dataset, this method improved the accuracy, recall, and F1 scores by an average of 5.3%, 5.6%, and 5.8%, respectively. In the ablation experiments of the DSDCN module, the classification accuracy of the model with the addition of the DSDCN module improved by 5.18%. In addition, the training method of transfer learning also led to an improvement in model classification performance, reflected in the classification accuracy, which increased by 7.4%. [ABSTRACT FROM AUTHOR]

In order to improve the accuracy of underwater object classification, according to the characteristics of sonar images, a classification method based on depthwise separable convolution feature fusion is proposed. Firstly, Markov segmentation is used to segment the highlight and shadow regions of the object to avoid the loss of information caused by simultaneous segmentation. Secondly, depthwise separable convolution is used to learn the deep information of images for feature extraction, which produces less network computation. Thirdly, features of highlight and shadow regions are fused by the parallel network structure, and pyramid pooling is added to extract the multi-scale information. Finally, the full connection layers are used to achieve object classification through the Softmax function. Experiments are conducted on simulated and real data. Results show that the method proposed in this paper achieve superior performance compared with other models, and it also has certain flexibility. [ABSTRACT FROM AUTHOR]

Epithelial protein lost in neoplasm (EPLIN) has been implicated as a suppressor of cancer progression. The current study explored EPLIN expression in clinical gastric cancer and its association with chemotherapy resistance. EPLIN transcript expression, in conjunction with patient clinicopathological information and responsiveness to neoadjuvant chemotherapy (NAC), was explored in two gastric cancer cohorts collected from the Beijing Cancer Hospital. Kaplan-Meier survival analysis was undertaken to explore EPLIN association with patient survival. Reduced EPLIN expression was associated with significant or near significant reductions of overall, disease-free, first progression or post-progression survival in the larger host cohort and Kaplan Meier plotter datasets. In the larger cohort EPLIN expression was significantly higher in the combined T1 + T2 gastric cancer group compared to the T3 + T4 group and identified to be an independent prognostic factor of disease-free survival and overall survival by multivariate analysis. In the smaller, NAC cohort, EPLIN expression was found to be significantly lower in tumour tissues than in paratumour tissues. EPLIN expression was significantly associated with responsiveness to chemotherapy which contributes to overall survival. Together, EPLIN appears to be a prognostic factor and may be associated with patient sensitivity to NAC. [ABSTRACT FROM AUTHOR]