Papers by Author: Jun Chen

Abstract: The paper proposes a novel discriminant non-negative matrix factorization algorithm and applies it to facial expression recognition. Unlike traditional non-negative matrix factorization algorithms, the algorithm adds discriminant constraints in low-dimensional weights. The experiments on facial expression recognition indicate that the algorithm enhances the discrimination capability of low-dimensional features and achieves better performance than other non-negative matrix factorization algorithms.

Abstract: In the article a new concurrency control protocol for real-time database (RTCC) is proposed. The protocol based on the traditional speculative concurrency control protocol (SCC). It dynamically establish the maximum of shadow to reasonably use the resources of system and add the quasi-commit phase to avoid many unnecessary restarting and enhance the concurrency of transaction. The theory of Petri net proof and results of experiment show that this protocol is feasible and effective , and it can meet the needs of real-time transaction. Speculative Concurrency Control is suitable especially for real-time database[1]. On the one hand, SCC is similar to PCC, to detect potentially harmful conflict as early as possible, so as to start a replacement schedule, thereby increasing the opportunities for transactions to meet time limits. On the other hand, SCC is similar to OCC, it allows the concurrency of conflict transaction and, therefore, avoids unnecessary matters that may affect the timely submission of delay. As a result, it alleviates the congestion problems and PCC restart OCC's problems, in order to better meet the transaction deadline. Based on the fact above mentioned, we propose a suitable real-time database concurrency control protocol [2,3] (RTCC: Concurrency Control protocol for Real-Time). The protocol of the SCC protocol on the traditional basis with a series of improvements can avoid a large number of unnecessary transaction restart and improve concurrency. Finally, by Petri net theory it can be proved feasible and correct.

Abstract: Trusted computing technology is a basic and entirely solution for security problems of computer. In order to solve the security problem caused by dynamics and the uncertainty of the grid environment, the article has put forward a new trust model based on behavior and trusted computing to deal with the trusting relationship among the entities, and take different methods to deal with to the trusting relationship between inner-domain and intra-domain based on the concept of Identity right, and a method for measuring and evaluating the behavior trust are provided.