Traditional time series imputation methods typically assume that time series data is derived from a linear dynamic system. However, the real-world time series show more non-linear characteristics. Therefore, a time series imputation model based on Long Short-Term Memory (LSTM) network with residual connection, called RSI-LSTM (ReSidual Imputation Long-Short Term Memory), was proposed to capture the non-linear dynamic characteristics of time series effectively and mine the potential relation between missing data and recent non-missing data. Specifically, the LSTM network was used to model the underlying non-linear dynamic characteristics of time series, meanwhile, the residual connection was introduced to mine the connection between the historical values and the missing value to improve the imputation capability of the model. Firstly, RSI-LSTM was applied to impute the missing data of the univariate daily power supply dataset, and then on the power load dataset of the 9th Electrical Engineering Mathematical Modeling Competition problem A, the meteorological factors were introduced as the multivariate input of RSI-LSTM to improve the imputation performance of the model on missing value in the time series. Furthermore, two general multivariate time series datasets were used to verify the missing value imputation ability of the model. Experimental results show that compared with LSTM, RSI-LSTM can obtain better imputation performance, and has the Mean Square Error (MSE) 10% lower than LSTM generally on both univariate and multivariate datasets.

In recent years, the short-term power load prediction model built with Recurrent Neural Network (RNN) as main part has achieved excellent performance in short-term power load forecasting. However, RNN cannot effectively capture the multi-scale temporal features in short-term power load data, making it difficult to further improve the load forecasting accuracy. To capture the multi-scale temporal features in short-term power load data, a short-term power load prediction model based on Multi-scale Skip Deep Long Short-Term Memory (MSD-LSTM) was proposed. Specifically, a forecasting model was built with LSTM (Long Short-Term Memory) as main part, which was able to better capture long short-term temporal dependencies, thereby alleviating the problem that important information is easily lost when encountering the long time series. Furthermore, a multi-layer LSTM architecture was adopted and different skip connection numbers were set for the layers, enabling different layers of MSD-LSTM can capture the features with different time scales. Finally, a fully connected layer was introduced to fuse the multi-scale temporal features extracted by different layers, and the obtained fusion feature was used to perform the short-term power load prediction. Experimental results show that compared with LSTM, MSD-LSTM achieves lower Mean Square Error (MSE) with the reduction of 10% in general. It can be seen that MSD-LSTM can better capture multi-scale temporal features in short-term power load data, thereby improving the accuracy of short-term power load forecasting.

The texture images obtained by multiple sensors of 3D color scanner have color difference, resulting in color block in the 3D color model surface. In order to solve this problem, a modified method based on color transfer was proposed. First, the comprehensive assessment quality function was used to choose the best one of the color texture images obtained by multiple sensors as the standard image. Then, the mean and variance of other texture images in each color channel were adjusted refering to the standard image. The proposed method was applied to texture image color correction of 3D human body color scanner. The result shows that, after modifying the color difference between texture images, the color block of the color 3D body model is significantly improved with more balanced and natural color. Compared with the classical method, the improved color transformation method and the method based on the minimum angle selection method, the subjective and objective evaluation results prove the superiority of the proposed method.

The algorithm of extracting pedestrian features based on texture information has the problems of redundant feature information and being unable to depict the human visual sensitivity, an algorithm named SF-LBP was proposed to extract pedestrian texture feature by Significant Local Binary Pattern which fuses the characteristics of human visual pedestrian system. Firstly, the algorithm calculated the significant factor in each region by saliency detection method. Then, it rebuilt the eigenvector of the image by significant factor weight and pedestrian texture feature, and generated the feature histogram according to local feature. Finally it integrated adaptive AdaBoost classifier to construct pedestrian detection system. The experimental results on INRIA database show that the SF-LBP feature achieves a detection rate of 97% and about 2%-3% higher than HOG (Histogram of Oriented Gradients) feature and Haar feature. It reaches recall rate of 90% and 2% higher than other features. It indicates that the SF-LBP feature can effectively describe the texture characteristics of pedestrians, and improve the accuracy of the pedestrian detection system.

Concerning the low efficiency of network transmission caused by redundant traffic, an algorithm named Packet Feature based Redundancy Traffic Elimination (PFRTE) was proposed based on the protocol-independent traffic redundancy elimination technique. Based on the grouping of packet size, PFRTE dynamically analyzed statistical bimodal characteristics and packet features of network traffic and regarded the size of the packet with the greatest capability of redundancy elimination as the threshold. It decided the boundary points by using sliding window method and calculated the fingerprint of block data within two boundary points. PFRTE encoded the redundant blocks in a simple way and transfered the encoded data instead of redundant data. The experimental results show that, compared with redundant traffic elimination algorithm based on maximum selection and static lookup table selection, PFRTE has the advantage of analyzing the redundancy statistics of network traffic dynamically, and the CPU consumption reduces both at server and client. Meanwhile, the algorithm is also effective with rate of redundancy elimination bytes saving of 8%-40%.

Researches on the audio classification have involved various types of vector features. However, multi-semantics of audio information not only have their own properties, but also have some correlations among them. Whereas, to a certain extent, the simple vector representation cannot represent the multi-semantics and ignore their relations. Tensor Uniform Content Locator (TUCL) was brought forward to express the semantic information of audio, and a three-order Tensor Semantic Space (TSS) was constructed according to the semantic tensor. Tensor Semantic Dispersion (TSD) can aggregate some audio resources with the same semantics, and at the same time, the automatic audio classification can be accomplished by calculating their TSD. And Radical Basis Function Tensor Neural Network (RBFTNN) was constructed and used to train intelligent learning model. For the problem of multi-semantic audio classification, the experimental results show that our method can significantly improve the classification precision in comparison with the typical method of Gaussian Mixture Model (GMM), and the classification precision of RBFTNN model is obviously better than that of Support Vector Machine (SVM).