Static program analysis is a cornerstone of modern software engineering - it is used to detect bugs and security vulnerabilities early before software is deployed. While there is a large body of research into the scalability and the precision of static analysis, the (un) soundness of static analysis is a critical issue that has not attracted the same level of attention by the research community. In this paper we investigate the question whether information harvested from stack traces obtained from the GitHub issue tracker and Stack Overflow Q&A forums can be used in order to complement statically built call graphs. For this purpose, we extract reflective call graph edges from parsed stack traces, and check whether these edges are correctly computed by Doop, a widely used tool for static analysis with built-in support for reflection analysis. We do find edges that Doop misses when analysing real-world programs, even when reflection analysis is enabled. This suggests that mining techniques are a useful tool to test and improve the soundness of static analysis.
Deep learning-based algorithms have been widely used in the low-dose CT imaging field, and have achieved promising results. However, most of these algorithms only consider the information of the desired CT image itself, ignoring the external information that can help improve the imaging performance. Therefore, in this study, we present a convolutional neural network for low-dose CT reconstruction with non-local texture learning (NTL-CNN) approach. Specifically, different from the traditional network in CT imaging, the presented NTL- CNN approach takes into consideration the non-local features within the adjacent slices in 3D CT images. Then, both low-dose target CT images and the non-local features feed into the residual network to produce desired high-quality CT images. Real patient datasets are used to evaluate the performance of the presented NTL-CNN. The corresponding experiment results demonstrate that the presented NTL-CNN approach can obtain better CT images compared with the competing approaches, in terms of noise-induced artifacts reduction and structure details preservation.