.
Semantic scene segmentation has become an important application in computer vision and is an essential part of intelligent transportation systems for complete scene understanding of the surrounding environment. Several methods based on convolutional neural networks have emerged, but they have some problems, including small-scale target loss, inaccurate detailed region segmentation, and boundary category confusion. Using shallow features, we exploit the capabilities of global context information according to the theory of pyramids. A weighted pyramid feature fusion module is constructed to fuse the feature maps of different scales generated by the backbone network, and the proportion of feature fusion is dynamically updated by trainable parameters. After that, a self-attention mechanism is introduced to discover information about spatial channel interdependencies. Finally, the atrous spatial pyramid pooling module of the DeepLabv3+ network is improved by connecting the atrous convolution with different dilation rates at the receptive field. The experimental results show 4.1% mean pixel accuracy and 3.92% mean intersection over union improvements in the proposed method compared with the DeepLabv3+, and the result of semantic segmentation is more accurate.
Abstract.As an important differentiated service model, proportional delay differentiation (PDD) aims to maintain the queuing delay ratio between different classes of requests or packets according to pre-specified parameters. This paper considers providing PDD service in web application servers through feedback control-based database connection management. To achieve this goal, an approximate linear time-invariant model of the database connection pool (DBCP) is identified experimentally and used to design a proportional-integral (PI) controller. Periodically the controller is invoked to calculate and adjust the probabilities for different classes of dynamic requests to use database connections, according to the error between the measured delay ratio and the reference value. Three kinds of workloads, which follow deterministic, uniform and heavy-tailed distributions respectively, are designed to evaluate the performance of the closed-loop system. Experiment results indicate that, the controller is effective in handling varying workloads, and PDD can be achieved in the DBCP even if the number of concurrent dynamic requests changes abruptly under different kinds of workloads.
The osteosarcoma (OS) microenvironment is composed of tumor cells, immune cells, and stromal tissue and is emerging as a pivotal player in OS development and progression. Thus, microenvironment-targeted strategies are urgently needed to improve OS treatment outcomes. Using principal component analysis (PCA), we systematically examined the tumor microenvironment (TME) and immune cell infiltration of 88 OS cases and constructed a TME scoring system based on the TMEscore high and TMEscore low phenotypes. Our analysis revealed that TMEscore high correlates with longer survival in OS patients, elevated immune cell infiltration, increased immune checkpoints, and increased sensitivity to chemotherapy. TMEscore low strongly correlated with immune exclusion. These observations were externally validated using a GEO dataset (GSE21257) from 53 OS patients. Our laboratory data also proved our findings. This finding enhances our understanding of the immunological landscape in OS and may uncover novel targeted therapeutic strategies.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.