Search

Person re-identification (re-ID) can be regarded as a retrieval problem. The challenge of this task is mainly lying in that people's appearance often undergoes dramatic changes across camera views due to severe occlusions, illumination, complex background clutter, and large pose variation. Recently, many deep learning approaches are successfully employed in person re-ID, which have achieved state-of-the-art performance in a period. However, it easily neglects the local discriminative details on the images and is not enough to cover discriminative information. Therefore, we employ ResNet50 as a basic network and propose a multi-loss siamese network with pyramid pooling structure for person re-ID. Our network can get a better discriminative representation and achieve a better performance on two large public datasets including Market1501 and CUHK03. We also report competitive accuracy compared with previous approach, which demonstrates the effectiveness of proposed method.

This paper presents a lighter protocol, and it removes the multicast burdens from RSVP to adapt to unicast applications. At the same time, when RSVP is used in wireless networks, some issues about mobility raise popular concerns. The proposed protocol a lightweight mobile RSVP protocol, solves the problems by the following mechanisms: changeless flow identifier, a new state management and “refresh” mechanism.

A main shortcoming of mobile Ad-hoc network’s reactive routing protocols is the large volume of far-reaching control traffic required to support the route discovery (RD) and route repair (RR) mechanism. Using a random mobility model, this paper derives the probability equation of the relative distance (RDIS) between any two mobile hosts in an ad-hoc network. Consequently, combining with average equivalent hop distance (AEHD), a host can estimate the routing hops between itself and any destination host each time the RD/RR procedure is triggered, and reduce the flooding area of RD/RR messages. Simulation results show that this optimized route repair (ORR) algorithm can significantly decrease the communication overhead of RR process by about 35%.

The Simple Mail Transfer Protocol (SMTP) based E-mail system is the most popular application in today's Internet world, but there are many dark clouds at the edge of the sunny sky. The two most well known threats are the "mail SPAM" and the "mail flooding attack". Mobile agents (MAs) are software agents that can transport themselves from a client computer to various servers for remote execution. Their mobility, autonomy and distributed intelligence are very friendly to today and tomorrow's Internet computation. In this paper, we introduce a prototype of the Internet mail transfer and check system based on intelligence mobile agents. This new system model can improve the current simple mail transfer protocol model to defend the mail SPAM and the mail flooding attack. Compatible with the traditional SMTP is the most important feature of our new model. Finally, we summarize some difficulties for turning this prototype model into practical software.