Based on an analysis of the problems of multiple interests of the user and multiple contents of the item existing in the collaborative filtering recommendation system, an improved item-based collaborative filtering algorithm is proposed. This new algorithm takes synthetically into account the influence of item attributes and user ratings. Experimental results indicate that the algorithm can satisfactorily solve the problems of multiple interests of the user and multiple contents of the item and provide better recommendation results even if the user ratings are very sparse.
This paper proposes an image classification accuracy prediction based unsupervised band selection method for hyper spectral image classification. The key of this method is the prediction of overall classification accuracy for each spectral band with no ground truth or training samples. Under the hypothesis of Gaussian Mixture Model (GMM), we build the explicit expression between the overall accuracy and the distribution parameters of each class, which is denoted as the overall accuracy prediction equation (OCPE). Then, by employing the unsupervised mixture models learning algorithm to predict these distribution parameters, the overall accuracy is computable on the basis of the OCPE. Then, the candidate band subset is obtained by selecting the bands with relatively high overall accuracy. Finally, we use the divergence based band decor relation algorithm to further remove the redundant bands. Real hyper spectral images based experiments show that our band selection method is effective in comparison with other three well-known unsupervised band selection techniques.
Taking Poyang Lake region as a case study area,this article analyzed the changes of tempo-spatial characteristics in land use based on Remote Sensing(RS),Geographical information system(GIS),correlation analysis method and land use change dynamic models.The result indicated that the simple land use change dynamic index of grassland and cultivated land are the largest.Approximately,0.6% of grassland and 0.3% of cultivated land changed every year.Water body and unused land are the most stable land uses.Only 0.05% of them changed every year.In terms of integrated land use dynamic index,Nanchang County is placed first where 4.9% of land use changed every year,and Jinxian County is placed the last where only 0.046% of land use changed every year.The results of correlation analysis showed that GDP and total financial income are the most visible factors of land use change.Affected by land use policy such as that of returning cultivated land to lake or to forest,land use level of 2004 in Poyang Lake region decreased slightly compared to that of 1992.
In the area of optical remote sensing image processing, object detection is among the utmost essential and difficult tasks. By virtue of the excellent feature representation abilities of deep convolutional neural networks (DCNNs), the performance of remote sensing object detection has recently increased dramatically. However, DCNN-based methods necessitate sufficient labeled training samples and tend to experience a considerable performance fall when training examples are insufficient. Many of the recently developed few-shot object detection (FSOD) methods attempt to solve the issue by using the idea of meta-learning, which intends to extract knowledge that can be generalized across various tasks. Despite its success, the negligence of the knowledge learned in the past, and inter-class correlations hinder the detection ability of novel classes. In this article, we propose a new meta-memory based FSOD approach named MM-RCNN. Specifically, MM-RCNN adopts a memory moduel to store each category's knowledge learned in the training stage and the memory-based external attention (MEA) to aggregate all categories' information simultaneously. Based on MEA, we design two feature enhancement modules for the region proposal network (RPN) and detection head to boost the performance. Experiments over two remote sensing benchmarks, i.e., DIOR and NWPU VHR10, verify the capability of our method (0.239 and 0.557 average mAP across all settings).
The technology of hyperspectral remote sensing, as an advanced technology in remote sensing, has been found wide application in many fields. However, the massive and high dimension data produce a challenge during its processing and analysis. Hyperspectral image fusion is rising as a new method which results from this background. The fused image would have enhanced information which is more understandable and decipherable for object recognition accurately. In this paper, we propose a novel method for image fusion and enhancement, using Empirical Mode Decomposition (EMD). EMD is a new data analysis method which expresses the tendency of signals at different scales by decomposing any complicated signal into a set of Intrinsic Mode Functions (IMFs). In this method, we decompose images from different hyperspectral band into their IMFs, and perform image fusion at the decomposition level. Based on an empirical understanding of the nature of the IMF, we devise adaptive weighting schemes which emphasize features from different band image, thereby increasing the information and visual content of the fused image.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)
