Biblio

With the increasing application of micro-nano satellite network, it is extremely vulnerable to the influence of internal malicious nodes in the practical application process. However, currently micro-nano satellite network still lacks effective means of routing security protection. In order to solve this problem, combining with the characteristics of limited energy and computing capacity of micro-nano satellite nodes, this research proposes a secure routing algorithm based on trust value. First, the trust value of the computing node is synthesized, and then the routing path is generated by combining the trust value of the node with the AODV routing algorithm. Simulation results show that the proposed MNS-AODV routing algorithm can effectively resist the influence of internal malicious nodes on data transmission, and it can reduce the packet loss rate and average energy consumption.

Facial recognition and detection is a traditional problem of computer vision, and the problem of facial key points detection is one of the most important branches of it. As the development of deep learning, more and more methods based on it are proposed for solving related issues, which bring numerous revolutionary changes. In our experiment, we propose a simplified method based on Convolution Neural Network for solving the problem of detecting 5 human facial key points. The method mainly consists of 2 sections. The first one is determining the facial area in an image, the output of which is a matrix representing coordinates of the facial area. The second one detects the relative position within the facial area and fine-tune it repeatedly. In the two sections, we design the structure of Neural Network for it, which size of hidden layers is small but stay efficient. We use Caffe, a popular open source framework of deep learning, to build our neural network and get the satisfactory result.