No abstract

COVID-19 is reshaping the relationship between humans and their living environment, potentially generating a profound impact on human physical and mental health and quality of life. The objective of this study was to explore how outdoor activities and the campus landscape impacted the subjective well-being (SWB) of Chinese university students in the pre-COVID-19 era (before December 2019) and during the COVID-19 era (from December 2019 to early December 2022). This study collected 439 valid questionnaires from students at a Chinese university, with the questions focusing on the frequency and length of time that university students of different genders, grades, and abilities to cover their living expenses participated in outdoor activities in the pre- and COVID-19 eras, as well as the changes in their SWB. Paired sample t tests revealed a significant decrease in SWB during the COVID-19 era compared with the pre-COVID-19 era, while independent sample t tests and a one-way ANOVA showed that changes in the SWB of university students pre- and during COVID-19 were not related to their gender or their ability to cover their living expenses, but were related to their grades. Hierarchical linear regression showed that outdoor activities in the COVID-19 era had a significant effect on mitigating the decline in students’ SWB. These results suggest a proactive way to increase resilience to future public health challenges and other crises of human life safety.

Few-shot segmentation focuses on the generalization of models to segment unseen object with limited annotated samples. However, existing approaches still face two main challenges. First, huge feature distinction between support and query images causes knowledge transferring barrier, which harms the segmentation performance. Second, limited support prototypes cannot adequately represent features of support objects, hard to guide high-quality query segmentation. To deal with the above two issues, we propose self-distillation embedded supervised affinity attention model to improve the performance of fewshot segmentation task. Specifically, the self-distillation guided prototype module uses self-distillation to align the features of support and query. The supervised affinity attention module generates high-quality query attention map to provide sufficient object information. Extensive experiments prove that our model significantly improves the performance compared to existing methods. Comprehensive ablation experiments and visualization studies also show the significant effect of our method on fewshot segmentation task. On COCO-20 i dataset, we achieve new state-of-the-art results. Training code and pretrained models are available at https://github.com/cv516Buaa/SD-AANet.