In order to solve the problem of uneven distribution of crowd and massive network learning parameters, a method for accurate high-density crowd counting was proposed, which is composed of Pixel-level Attention Mechanism (PAM) and improved single-column crowd density estimation network. First of all, the PAM was used to generate a high-quality local crowd density map by classifying the crowd images at pixel level, and the Full Convolutional Network (FCN) was used to generate the density mask of each image, and the pixels in image were divided into different density levels. Then, using the generated density mask as the label, the single-column crowd density estimation network was used to learn more representative features with fewer parameters. Before this method was proposed, the counting error of Network for Congested Scene Recognition (CSRNet) method was the smallest on part_B of Shanghaitech dataset, the UCF_CC_50 dataset and the WorldExpo'10 dataset. Comparing the error results of proposed method with CSRNet, it is found that this method has the Mean Absolute Error (MAE) and Mean Squared Error (MSE) on part_B of Shanghaitech dataset reduced by 8.49% and 4.37%; the MAE and MSE on UCF_CC_50 dataset decreased by 58.38% and 51.98% respectively, which are of significant optimization, and the MAE of overall average value on the WorldExpo'10 dataset reduced by 1.16%. The experimental results show that when counting the unevenly distributed high-density crowd, the method of combination of PAM and single-column crowd density estimation network can effectively improve the accuracy and training efficiency of high-density crowd counting.

To detect building-damage areas in post-seismic high-resolution remote sensing images, a building-damage detection method based on multi-features was proposed. Firstly, Morphological Attribute Profile (MAP) and Local Binary Pattern (LBP) operator were used to extract geometric features and texture features. Then, Random Forest (RF) classifier was applied to extract damaged building regions so as to form the preliminary results. At last, for segmented objects, the ultimate building-damage area was obtained by computing the damaged ratio of each object. Experiments were carried out on Yushu post-seismic aerial remote sensing images whose spatial resolution was 0.1 m. Results show that this method improves overall accuracy by 12% compared with Morphological Profile (MP)-based method. The results indicate that the proposed method can effectively detect building-damage areas with high accuracy in post-seismic high-resolution images.

A multi-person decision method based on the grey additive linguistic variables weighted aggregation operator is presented to solve the Network Security Risk Evaluation problems in mobile bank, in which the attribute values take the form of the grey additive linguistic variables(GALV). Firstly, some properties are defined, such as the concept and the relational calculation rules of grey additive linguistic variables. Then, some operators will be defined to solve the Network Security Risk Evaluation problems in mobile banking, such as grey additive linguistic weighted aggregation operator, and grey additive linguistic ordered weighted aggregation operator. At last, an example involved in mobile bank shows the effectiveness of this method.

Considering the correlation and dependence among indicators, a multi-person decision model based on analytic network process was designed. In this model, the network security evaluation index was elicited by means of analytic network process, and then, a bi-level programming based on weighted Euclidean distance was used to synthesize the individual decision-making results. An illustrative example was given to demonstrate the feasibility and validity of the proposed method. The result shows that the proposed model has higher credibility to evaluate the network security in commercial banks.