The widespread use of smartphones has made ridesharing a popular transportation option, connecting passengers with drivers seamlessly. However, existing ridesharing models raise privacy concerns as users must disclose their travel plans without assurance of a match. Previous privacy solutions are either impractical or insecure. To address this, we propose a novel privacy-preserving ridesharing mechanism based on probabilistic matching. Our approach ensures robust privacy protection for both drivers and riders while maintaining efficiency.

Location Based Services (LBSs) have become increasingly popular in the past decade, allowing mobile users to access location-dependent information and services. To protect user privacy while using LBSs, various Location Privacy Protection Mechanisms (LPPMs) have been proposed that obfuscate users' true locations through random perturbation. However, adversaries can still exploit the temporal correlation between a user's locations in multiple LBS queries to improve inference accuracy. In this paper, we introduce a novel location inference attack that strikes a good balance between inference accuracy and computational complexity by effectively exploiting temporal correlation. Simulation studies using synthetic and real datasets confirm the advantages of our proposed attack.

A large number of clinical studies have proved that body shape is closely related to health. Considering the actual situation, the traditional health examination has some limitations, such as low efficiency, complex process and unable to achieve real-time monitoring. In order to make health prevention more targeted and reduce the risk and deterioration of some diseases, we propose a prediction model of obesity diseases. In this model, a method based on three-dimensional human body model is used to obtain the size of human body. The shape features of human body are extracted from two human body images (front image and side image), and the dense ellipse model proposed by us is used to measure the size of human body on the three-dimensional human body model. We also summarize the traditional and nontraditional anthropometric parameters for predicting obesity diseases, and give the possible disease types and risks according to the measured body size. Users only need to take two photos through smart phones, and about 15 seconds can get health reports including anthropometric dimensions and possible types of diseases. In summary, our method can measure 38 body dimensions with an accuracy rate of 97.4%, calculate 15 anthropometric parameters related to obesity, and analyze 10 possible types of diseases.

There are many illegal memory access (IMA) defects in C programs, for example, null pointer dereference, buffer overflow and array out of bounds. When C programs are running, these defects may cause software failure, and may lead to serious consequences. In order to fully detect these IMA defects, we apply abstract regions to simulate memory blocks that allocated to memory objects at the runtime, and use the region-based model to describe memory states of each program point based on the result of dataflow analysis. Based on the region-based model, the determining criterion of IMA defects is introduced for detecting these defects. The method has been implemented in the defect detection tool DTSC, and 5 open source projects are tested, the test results show the validity of our method.