Large Feature Mining With Ensemble Learning for Image Forgery Detection

The detection of different types of forgery manipulation including seam-carving in JPEG images is a hot spot in image forensics. Seam carving was originally designed for content-aware image resizing. It is also being used for forgery manipulation. It is still very challenging to effectively identify the seam carving forgery under recompression. To address the highly challenging detection problems, this chapter introduces an effective approach with large feature mining. Ensemble learning is used to deal with the high dimensionality and to avoid overfitting that may occur with some traditional learning classifier for the detection. The experimental results validate the efficacy of proposed approach to detecting JPEG double compression and exposing the seam-carving forgery while the JPEG recompression is proceeded at the same quality and a lower quality, which is generally much harder for traditional detection methods. The methodology introduced in this chapter provides a strategy and realistic approach to resolve the highly challenging problems in image forensics.

Research on multi-label classification method of transformer based on DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm

With the improvement of the intelligent level of power grid and the enhancement of the integrated characteristics of power grid, the degree of discretization of massive data of power equipment gradually increases, which brings great challenges to the safe and stable operation of power grid. How to process and analyze data effectively has become an important research content. Transformer is an important electrical equipment, therefore it is of great significance to monitor the operation status of transformer, to construct transformer operation characteristic label system based on multi-source heterogeneous data, and to realize multi-label classification function. In this paper, a transformer multi-label classification method of transformer based on DBSCAN(Density-Based Spatial Clustering of Applications with Noise) clustering algorithm is proposed, which can accurately identify outliers as Noise without input of the number of clustering to be divided, realize the key feature mining of transformer state, and to realize to provide flexible information association and historical data for dispatch and control operators.

Evaluation Method of Chemical Technology Safety Practical Operation Ability Based on Stochastic Model

The current evaluation method does not consider the coordination and overall situation of various practical skills, which affects the evaluation contact degree and leads to hidden dangers of production safety. Therefore, an evaluation method of chemical technology safety practical operation ability based on a stochastic model is proposed. Using association rules to design the feature mining algorithm of practical operation ability, combined with crosslayer coding strategy, the mapping relationship between safety accidents and practical operation ability is established. The chemical safety practical operation skills are simulated by a random model, and the practical operation skills are classified and coordinated according to the simulation predictive control results. According to the coordination results, the evaluation index system is constructed, the index weight is determined, and the evaluation model of chemical technology safety operation ability is constructed. The experimental results show that the maximum connection degree of the evaluation method designed in this paper is 88.06%, and the evaluation results are more accurate, which is helpful to improve the production safety of chemical enterprises.

Feature mining method of information system operation and maintenance big data based on window function

In this paper, a window function based feature mining method for the operation and maintenance of big data in information system is proposed. The time clustering feature vector is combined with window function to reduce the dimension of operation and maintenance data of high-dimensional information system. The operation and maintenance data feature subset is segmented according to the similar feature level, and the redundant features of operation and maintenance data are removed to complete the information system operation and maintenance big data feature mining. The simulation results show that the proposed method has better clustering effect, fewer iterations and shorter mining time.

Learning Hierarchical Representations with Spike-and-Slab Inception Network

Recently, deep convolutional neural networks (CNN) with inception modules have attracted much attention due to their excellent performances on diverse domains. Nevertheless, the basic CNN can only capture a univariate feature, which is essentially linear. It leads to a weak ability in feature expression, further resulting in insufficient feature mining. In view of this issue, researchers incessantly deepened the network, bringing parameter redundancy and model over-fitting. Hence, whether we can employ this efficient deep neural network architecture to improve CNN and enhance the capacity of image recognition task still remains unknown. In this paper, we introduce spike-and-slab units to the modified inception module, enabling our model to capture dual latent variables and the average and covariance information. This operation further enhances the robustness of our model to variations of image intensity without increasing the model parameters. The results of several tasks demonstrated that dual variable operations can be well-integrated into inception modules, and excellent results have been achieved.