The existing fall detection works mainly focus on indoor scenes， and most of them only model people’s body posture features， ignoring background information of the scene and the interaction information between people and the ground. Aiming at the problem， from the perspective of practical application of elevator scene， a fall detection algorithm based on scene prior and attention guidance was proposed. Firstly， elevator historical data was used to automatically learn the scene prior information from people’s trajectories by Gaussian probability distribution modelling. Then， the scene information was taken as a spatial attention mask and fused with the global features of the neural network to focus on local information of the ground area. After that， the fused local and global features were further aggregated using adaptive weighting method to improve the robustness and discriminative ability of the generated features. Finally， the features were fed into a classifier module consisting of a global average pooling layer and a fully connected layer to perform the fall prediction and classification. Experimental results show that the detection accuracy of the proposed algorithm on the self-built elevator scene dataset Elevator Fall Detection Dataset and the public UR Fall Detection Dataset reached 95.36% and 99.01% respectively， which is increased by 3.52 percentage points and 0.61 percentage points respectively compared with that of ResNet50 with complicated network structure. It can be seen that proposed attention mechanism with Gaussian scene prior guidance can make the network focus on information of the ground area， which is more conducive to detect fall events. By using it， the detection model has high accuracy， and the algorithm meets the real-time application requirements.

Concerning the situation that the ecological security has a lot of uncertaint factors and is dynamic, the Set Pair Analysis (SPA) theory and Markov chain were combined for dynamic assessment of ecological security, and a dynamic assessment method of ecological security was proposed combining the state assessment and trend analysis. The method adopted connection degree of SPA to represent uncertain knowledge, used the connection number of SPA to classify ecological security level, and built the comprehensive evaluation model of the system. Through the analysis of the development trend of ecological security, using the ergodicity of Markov chain and statistical quality of Monte Carlo method, the state of ecological security in the next moment could be predicted. This method uses limited assessment data and historical data, dynamically assesses the development and changes of ecological security, and provides a theoretical basis for the safety management.

Currently, there is a lack of model design and development method for complicated behaviors in the research on model-based Web User Interface (UI), which severely limits the engineering application of the method. Based on Web UI design patterns, a formalized description language for Web interface behavior was proposed, which can be used to model complicated behaviors of Web UI. The automatic code generation for behavior models was also achieved consequently. The experimental results show that this method is greatly flexible in the development of UI design and improves the reusability of UI design patterns.

Traditional Markov Random Field (MRF) models have two inherent limitations that are low order property of pixel neighborhoods and selecting parameters by hand. In this paper, we adopted a new machine learning method of score matching and get a group of parameters of high order MRF models by learning from training image data. We demonstrated the capabilities of the learning MRF models by applying them to image denoising according to Bayesian rule. Imaging denoising experiments show that our denoising algorithm can produce excellent result in the Peak Signal-to-Noise Ratios (PSNR) and subjective visual effect. Thus, our learning method is effective.