Narrow Results

Traditional knowledge maps can only deal with single modal data. The combination of information fusion based multimodal knowledge maps and blockchain technology can effectively improve the expression ability and security of knowledge maps. This paper constructs a multimodal knowledge graph through information fusion technology, and combines blockchain technology to improve the data security, credibility and privacy protection capabilities of the knowledge graph, and further realizes the intelligent reasoning and efficient application of cross-domain knowledge to cope with complex and diverse practical application scenarios. This paper combined multimodal knowledge graph with blockchain technology, and conducted research on the construction efficiency, query efficiency, application effect of multimodal knowledge graph and application effect of multimodal knowledge graph fusion blockchain technology in the experimental section. It was known that the construction of a multimodal knowledge graph based on information fusion only took about 1h; the data query time of the multimodal knowledge graph based on information fusion was maintained at around 1s; the access time of distributed storage based on blockchain technology was maintained at around 1s; the cost of using blockchain technology for knowledge graph calculation was around 1,000yuan/month. The multimodal knowledge graph and blockchain technology based on information fusion can promote the development and application of knowledge graphs in various fields.

This paper proposes an intelligent 2ν-support vector machine based match score fusion algorithm to improve the performance of face and iris recognition by integrating the quality of images. The proposed algorithm applies redundant discrete wavelet transform to evaluate the underlying linear and non-linear features present in the image. A composite quality score is computed to determine the extent of smoothness, sharpness, noise, and other pertinent features present in each subband of the image. The match score and the corresponding quality score of an image are fused using 2ν-support vector machine to improve the verification performance. The proposed algorithm is experimentally validated using the FERET face database and the CASIA iris database. The verification performance and statistical evaluation show that the proposed algorithm outperforms existing fusion algorithms.

Several variants of spiking neural P systems (SNPS) have been presented in the literature to perform arithmetic operations. However, each of these variants was designed only for one specific arithmetic operation. In this paper, a complete arithmetic calculator implemented by SNPS is proposed. An application of the proposed calculator to information fusion is also proposed. The information fusion is implemented by integrating the following three elements: (1) an addition and subtraction SNPS already reported in the literature; (2) a modified multiplication and division SNPS; (3) a novel storage SNPS, i.e. a method based on SNPS is introduced to calculate basic probability assignment of an event. This is the first attempt to apply arithmetic operation SNPS to fuse multiple information. The effectiveness of the presented general arithmetic SNPS calculator is verified by means of several examples.

We use two traditional indices and propose a new fusion index to quantify the echo chirp signal amplification effectively. They are the cross-correlation coefficient, the spectral amplification factor and the fusion index. Through investigating different numerical examples, we verify the validation of the re-scaled vibrational resonance (VR) method for the echo chirp signal by using the two traditional indices. Besides, taking into account that a single index has its own shortness, we combine the advantages of the two traditional indices mentioned and propose a new fusion index based on the information fusion method. Then we verify the superiority of the fusion index. Furthermore, the effect of noise on VR is also investigated. Finally, with the help of the adaptive inertia weight particle swarm optimization (APSO) algorithm, we propose the adaptive VR in noise background and verify its validity.

Many real life problems are characterized by the structure of data derived from multiple sensors. The sensors may be independent, yet their information considers the same entities. Thus, there is a need to efficiently use the information rendered by numerous datasets emanating from different sensors. A novel methodology to deal with such problems is suggested in this work. Measures for evaluating probabilistic classification are used in a new efficient voting approach called "selective voting", which is designed to combine the classification of the models (sensor fusion). Using "selective voting", the number of sensors is decreased significantly while the performance of the integrated model's classification is increased. This method is compared to other methods designed for combining multiple models as well as demonstrated on a real-life problem from the field of human resources.

Aiming at the problems that the existing assessment methods are difficult to solve, such as the low efficiency and uncertainty of network security situation assessment in complex network environment, by constructing the characteristic elements of network security big data, a typical model based on deep learning, long short-term memory (LSTM), is established to assess the network security situation in time series. The hidden relationship and change trend of network security situation are automatically mined and analyzed through the deep learning algorithm of big data, which greatly improves the prediction accuracy of security situation. Experimental analysis shows that this method has a better assessment effect on network threats, has higher learning efficiency than the traditional network situation assessment methods, and has strong representation ability in the face of network threats. It can more accurately and effectively assess the changing trend of big data security situation in the future.

To improve the real-time performance and the target adaptability of penetration fuze detonation control systems, and to enhance the system fusion processing capability for multi-sensor information, this paper uses a modular design concept to construct a miniaturized (ø38mm×4mm) fuze detonation control system that is capable of real-time processing of data from multiple information sources. The core component of this system is the GD32E230 microcontroller, which features a high dominant frequency and low power consumption. This device is integrated with a ferroelectric memory and signal processing circuits that match the sensors. To address the issue of unclear traditional acceleration signal penetration and the difficulties associated with the identification of these signals, the approach in this paper improves feature recognition accuracy through rapid acquisition and fusion of multiple types of sensor output signal, and self-adaptive identification of multilayered targets and single-layer thick targets is achieved. During the programming of the embedded system, the hardware register is operated directly, the instruction execution sequence is optimized, and the program execution efficiency is improved by using the function characteristic that some microcontroller unit peripherals do not occupy the central processing unit when working, thus allowing the intended purpose of improving the system’s real-time performance to be achieved. A semi-physical simulation method is then used to verify the performance of the penetration fuze detonation control system. The results obtained show that the system has 100%-layer counting accuracy for multilayered targets and a relative error of less than 1% for the calculated residual velocities of single-layer thick targets, thus validating the effectiveness of the system.

In this research paper, we propose an automatic segmentation method of multispectral magnetic resonance image (MRI) of the human brain using an information fusion approach through the framework of the possibility theory. The fusion process is summarized into three essential steps. First, a data is extracted from the various images and modeled in a common mathematical framework, in this step the fuzzy C-means (FCM) algorithm is chosen. The combination rule is used to combine this information in the second step. A final segmented image is the result of the last phase. Our experimental results using simulated brain MRI datasets show that the proposed approach overcome the impact of the noise and substantially improve the accuracy of image segmentation.

This paper focuses on integrating information from RGB and thermal infrared modalities to perform RGB-T object tracking in the correlation filter framework. Our baseline tracker is Staple (Sum of Template and Pixel-wise LEarners), which combines complementary cues in the correlation filter framework with high efficiency. Given the input RGB and thermal videos, we utilize the baseline tracker due to its high performance in both of accuracy and speed. Different from previous correlation filter-based methods, we perform the fusion tracking at both the pixel-fusion and decision-fusion levels. Our tracker is robust to the dataset challenges, and due to the efficiency of FFT, our tracker can maintain high efficiency with superior performance. Extensive experiments on the RGBT234 dataset have demonstrated the effectiveness of our work.

This paper presents an efficient technique for unsupervised content-based segmentation in stereoscopic video sequences by appropriately combined different content descriptors in a hierarchical framework. Three main modules are involved in the proposed scheme; extraction of reliable depth information, image partition into color and depth regions and a constrained fusion algorithm of color segments using information derived from the depth map. In the first module, each stereo pair is analyzed and the disparity field and depth map are estimated. Occlusion detection and compensation are also applied for improving the depth map estimation. In the following phase, color and depth regions are created using a novel complexity-reducing multiresolution implementation of the Recursive Shortest Spanning Tree algorithm (M-RSST). While depth segments provide a coarse representation of the image content, color regions describe very accurately object boundaries. For this reason, in the final phase, a new segmentation fusion algorithm is employed which projects color segments onto depth segments. Experimental results are presented which exhibit the efficiency of the proposed scheme as content-based descriptor, even in case of images with complicated visual content.

This survey gives a review of recent artificial intelligence-related research directions that are considered priority areas by the U.S. Air Force and targeted for basic research funding by Air Force Office of Scientific Research. These research areas include space situational awareness, autonomous systems, sensing and information fusion, surveillance, navigation, robust decision making, human-computer interfaces, and computational and machine intelligence. The possible contributions of artificial intelligence to these topics will be described and illustrated whenever possible by recently awarded grants.

Fuzzy aggregation is the way in which different contributions to the same fuzzy fact are merged together to obtain a possibility distribution representative of the acquired knowledge. The choice of the aggregation function is a fundamental step in the definition of inference framework. In most cases aggregation has some monotonicity property and this can lead to saturation problems in complex frameworks, particularly in stateful rational agents. In this paper, we propose an extension to the fuzzy aggregation to handle these cases and apply fuzzy reasoning to complex KBs. We especially focus on Mamdani inference framework, where aggregation is implemented by a triangular conorm.

This paper introduces two new fusion rules for combining quantitative basic belief assignments. These rules although very simple have not been proposed in literature so far and could serve as useful alternatives because of their low computation cost with respect to the recent advanced Proportional Conflict Redistribution rules developed in the DSmT framework.

The combination of identical S-approximation spaces, except with different decider mappings, is studied in this paper by considering the construction of more complex S-approximation spaces from simpler ones that use different decision criteria, e.g., due to levels of expertise. It can be used to model group decision making problems where each decider makes an independent decision based on a shared knowledge map, e.g., several doctors with the same knowledge and different, independent decision criteria decide on a possible disease(s) for a patient, based on the same set of observations. These results can formalize the management of distributed uncertainty and can be used to invent novel distributed uncertain data processing algorithms. Also, we introduce the decider significance concept to minimize the number of combinations to obtain the same effect as the original combination. We show that finding a minimum set of significant deciders is NP-hard. and give an illustrative example in a medical expert system.

We describe and evaluate information fusion by fuzzy integration in a robust, high performance face recognition system. The system uses fuzzy integrals to combine classifiers operating at different image resolutions. Recognition is carried out by distance classification of transformed vectors of local autocorrelation coefficients. The transformation is determined by linear discriminant analysis. A large database of 11,600 images of 116 persons is used to determine the system performance. After being trained to recognize 60 persons, it is tested on images of all persons in the database. Both training and test stages use 50 images of each person. Under two different training schemes, it achieves peak recognition rates of 98.4% and 97.9%, respectively, accepting only 1.6% and 2.4% of the unknown faces. This exceeds the performance of any of the individual classifiers by at least 10%. Moreover, it exceeds earlier results obtained by multiple resolution averaging on the same database by at least 1.0%.

Fitting P-S-N curve with small-size sample of fatigue test data is significant in engineering applications. Although several small sample-based P-S-N curve fitting methods have been developed, complexity in mathematics and/or the unrealistic assumption of the methods hinder their application seriously. Based on the principle of probabilistically mapping from the probability distribution of specimen property to that of fatigue life of the specimen, this paper presents a new, easy to apply P-S-N curve fitting method. By collecting the life distribution information dispersed in several small-size samples of fatigue lives tested under different cyclic stress levels, a large-size sample of equivalent fatigue life data can be built based on the mapping mechanism as well as the uniqueness of the relationship between fatigue life standard deviation and cyclic stress level. The basic viewpoint is that the fatigue lives tested at any cyclic stress levels can be equivalently converted to an arbitrary baseline stress level according to the life distribution–stress relationship, and this principle can be applied to determine the P-S-N curves with a limited number of test data. Test results illustrate that the P-S-N curves obtained by such methods with 30, 24 or 20 samples, respectively, are close to those obtained by the conventional test method with 60 or 40 samples.

To support the retrieval, fusion and discovery of multimedia information, a spatial/temporal query language for multiple data sources is needed. In this paper we describe a spatial/temporal query language, the ∑QL, which is based upon the σ-operator sequence and in practice expressible in an SQL-like syntax. The general σ-operator and temporal σ-operator are explained, and applications of the σ-query language to vertical/horizontal reasoning and hypermapped virtual world are discussed.

In this paper, we address the problem of technical analysis information fusion in improving stock market index-level prediction. We present an approach for analyzing stock market price behavior based on different categories of technical analysis metrics and a multiple predictive system. Each category of technical analysis measures is used to characterize stock market price movements. The presented predictive system is based on an ensemble of neural networks (NN) coupled with particle swarm intelligence for parameter optimization where each single neural network is trained with a specific category of technical analysis measures. The experimental evaluation on three international stock market indices and three individual stocks show that the presented ensemble-based technical indicators fusion system significantly improves forecasting accuracy in comparison with single NN. Also, it outperforms the classical neural network trained with index-level lagged values and NN trained with stationary wavelet transform details and approximation coefficients. As a result, technical information fusion in NN ensemble architecture helps improving prediction accuracy.

Retinal diseases and systemic diseases, such as diabetic retinopathy (DR) and Alzheimer’s disease, may manifest themselves in the retina, changing the retinal oxygen saturation (${\mathrm{\text{SO}}}_2$) level or the retinal vascular structures. Recent studies explored the correlation of diseases with either retina vascular structures or ${\mathrm{\text{SO}}}_2$ level, but not both due to the lack of proper instrument or methodology. In this study, we applied a dual-modal fundus camera and developed a deep learning-based analysis method to simultaneously acquire and quantify the ${\mathrm{\text{SO}}}_2$ and vascular structures. Deep learning was used to automatically locate the optic discs and segment arterioles and venules of the blood vessels. We then sought to apply machine learning methods, such as random forest (RF) and support vector machine (SVM), to fuse the ${\mathrm{\text{SO}}}_2$ level and retinal vessel parameters as different features to discriminate against the disease from the healthy controls. We showed that the fusion of the functional (oxygen saturation) and structural (vascular parameters) features offers better performance to classify diseased and healthy subjects. For example, we gained a 13.8% and 2.0% increase in the accuracy with fusion using the RF and SVM to classify the nonproliferative DR and the healthy controls.

Two-sided matching decision-making (TSMDM) problems widely exist in human being’s life. In practical TSMDM problems, subjects who are being matched with usually tend to provide linguistic evaluations for convenience. Owing to the fuzziness of decision-making environment, subjects may provide several linguistic terms associated with their probabilities as the evaluations, which can be described as the probabilistic linguistic term sets (PLTSs). To model such scenarios, this paper proposes a TSMDM approach with probabilistic linguistic evaluations. The probabilistic linguistic evaluations are firstly normalized with the preservation of ignored probabilities. The normalized evaluations are then aggregated into the comprehensive evaluations by using the defined Choquet integral-based probabilistic linguistic aggregation (CIPLA) operator. Thereafter, the satisfaction degrees of two sides of subjects are calculated based on the closeness coefficient in TOPSIS method. On the basis of this, a multi-objective TSMDM model aiming to maximize the comprehensive satisfaction degree is built and transformed into a single-objective TSMDM model considering the weights of two sides. To solve the single-objective TSMDM model is to determine the optimal matching result. An illustrative example of matching enterprises and knowledge-sender universities is introduced to validate the proposed approach and compare it with other existing TSMDM approaches. The results illustrate that the proposed approach can not only avoid information loss, but also effectively integrate PLTSs with correlative criteria.