Zero-shot classification(ZSC) aims to classify images from the class whose training samples are unavailable. A typical method addressing this issue is to learn a projection from feature space to attribute space so tha...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Zero-shot classification(ZSC) aims to classify images from the class whose training samples are unavailable. A typical method addressing this issue is to learn a projection from feature space to attribute space so that a relation of training samples and test samples could be built. However, the projection merely learned from training samples does not apply in unseen classes due to domain shift between them. To tackle this issue, we propose a novel method in this paper that jointly learns coupled autoencoders to alleviate the distribution divergence of *** learn a projection by adopting encoder-decoder paradigm in both seen and unseen classes. The proposed method is evaluated for zero-shot recognition on two benchmark datasets, achieving competitive results.
Object detection is a very significant task for a huge range of applications. For example, the detection of military vehicles is very useful for the defense and intelligence. In recent years, hyperspectral imagery(HSI...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Object detection is a very significant task for a huge range of applications. For example, the detection of military vehicles is very useful for the defense and intelligence. In recent years, hyperspectral imagery(HSI) which is generated by remote sensing systems can provide tremendous information about the spectral characteristics. Due to this characteristic, object detection using HSI becomes hot research topic. In this paper, we propose a strategy for military object detection by extracting multiple information from HSI. Firstly, we generate the superpixels from HSI by principle component analysis(PCA) and k-means clustering. Then, self-similarity method is used to calculate the correlation between each superpixel and the object spectral. At last, the shape information is extracted from the masses which have high correlation value and is used to detect the specific military objectives. Results from HSI demonstrate the benefits of the proposed strategy regarding its effectiveness at detecting specific objectives.
For the sake of protecting users' privacy from the malicious data which is shared by other people and leasing the pressure of the clouds, an approach of verification based on the Blockchain is put forward in our a...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
For the sake of protecting users' privacy from the malicious data which is shared by other people and leasing the pressure of the clouds, an approach of verification based on the Blockchain is put forward in our article. By using the Blockchain to record the hash value and other necessary information of data sharing by other people, we can guarantee that the data user received from a third-party source(such as a cloud storage platform) is the original uploaded data indeed. After taking some experiments on the following method. We confirm that it can be easily distinguished whether a data has been modified by malicious people through our Blockchain-based approach. This Blockchain-based approach of verification can effectively help users find out if the data received is the one exactly he wants. In addition, our approach will imply that the data is not the original one and that data cannot be opened or executed.
Complicated optical scenes generally exist on the water surface. The noise originated from the light reflection seriously blocks the performance of the object detection *** paper analyzes the correlation between the i...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Complicated optical scenes generally exist on the water surface. The noise originated from the light reflection seriously blocks the performance of the object detection *** paper analyzes the correlation between the incident light angle and the flare noise, the relation between the solar altitude and the polarization state of the reflected light. Further, the correlation between the various imaging factors, such as the imaging time, the imaging angle of view, the polarization direction and the polarization state, is adjusted and optimized by the polarization imaging experiment. According to the experimental results, our polarization imaging method has good capability to suppress the noises generated by the light reflection, improving the accuracy of the object detection results.
Dynamic stochastic resonance(DSR) based dark and low-contrast image enhancement has attracted more and more attention in recent years. For DSR based image enhancement,noise is essential and will be enhanced simultaneo...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Dynamic stochastic resonance(DSR) based dark and low-contrast image enhancement has attracted more and more attention in recent years. For DSR based image enhancement,noise is essential and will be enhanced simultaneously with the contrast of the image, which is undesirable for improvement of perceptual quality. Nonlinear anisotropic diffusion(NAD)is one of the most widely used denoising methods due to good performance of edge preservation, but often fails for contaminated images with high level of noise. In this paper,we propose a novel partial differential equation method for image enhancement by introducing filtering into the stochastic resonance equation, and we consider two kinds of NAD *** results demonstrate that the improved methods can not only increase brightness and contrast of the dark and low-contrast images efficiently by optimum iterations, but also remove the noise while preserving edges well, and therefore can achieve good perceptual quality.
With the development of network technology, decision support and data mining has been widely applied in many fields, such as finance, marketing, quality analysis. But it isn't applied widely in the evaluation of n...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
With the development of network technology, decision support and data mining has been widely applied in many fields, such as finance, marketing, quality analysis. But it isn't applied widely in the evaluation of network education. In order to improve the quality of personal training, universities regard the "the quality of teaching and teaching reform project" as the starting point, carry out teaching and management reform and explore the new talent cultivation mode. This paper summarizes the relative theories and methods of decision support, then introduces association rule in data mining. First, we analyze the classical association rules mining Apriori algorithm, then analyse this algorithm's disadvantages and improvement ways. Second,we use the Apriori algorithm, which can evaluate the network education from different aspects, thus providing manager with clear and actionable feedback to enhance their practice. Finally,the algorithm's efficiency is assessed and we points out the further research problems.
With the rapid development of agent modeling technology, the modeling of agent behavior plays an important role in simulation system especially for games and training system. As the basic part of agent behavior modeli...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
With the rapid development of agent modeling technology, the modeling of agent behavior plays an important role in simulation system especially for games and training system. As the basic part of agent behavior modeling, the modeling of Tactical Position Selection(TPS) directly affects the the user experience. Traditional TPS formulation methods do not take team cooperation into consideration. The modeling of this factor is an important area for TPS research. This paper abstracts Team-TPS(TTPS) problem as the Constraint Satisfaction Problem(CSP) and proposes an improved synchronous backtracking algorithm based on entropy ordering to solve the issue. In order to verify the effectiveness of the proposed model, a simulation experiment was carried out. The results demonstrate that for the same tactical task, the proposed Team-TPS model can find better solutions and of high efficiency.
Aiming at the path planning problem of mobile robots in unknown environment,a random chemical reaction algorithm based on double containers is *** new algorithm is inspired by the bacterial foraging algorithm and appl...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
Aiming at the path planning problem of mobile robots in unknown environment,a random chemical reaction algorithm based on double containers is *** new algorithm is inspired by the bacterial foraging algorithm and applies a strategy based on randomly distributed chemical molecules around the robot,these molecules search the best path to the target location,while relying on the robot's own sensors to avoid *** selection criterion of the optimal path depends on the potential energy error and the distance error of the molecules on the detected obstacles or targets,and guides the robot to move in this *** order to evaluate the effectiveness of the algorithm,a series of simulations were made in three different scenarios and the results were *** results show that the algorithm can effectively avoid falling into local *** the same time,the algorithm proposed can provide a shorter and smoother path,and the operation is more simple.
The three-dimensional geological modeling and orebody visualization based on a small number of drilling cores are key issues for three-dimensional geological exploration. In this paper, we present an automatic pipelin...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
The three-dimensional geological modeling and orebody visualization based on a small number of drilling cores are key issues for three-dimensional geological exploration. In this paper, we present an automatic pipeline of delineation and 3 D profile mapping for potential exploration targets. As an integrated threedimensional modeling process, the framework includes three key modules such as the terrain generation, delineation and three-dimensional profile mapping. The delineation is achieved by calculating the cutoff grade isosurface with a novel curved spline model. For the threedimensional profile mapping, it is made by using the Kriging interpolation to generate the profile at an arbitrary angle within the delineation area. Finally, the proposed framework is validated by using the real mineral data of some drilling cores of Wulonggou area in Qinghai province of China, and the visualization results verify its effectiveness.
In this study, a hierarchical architecture for the intermittent control under the minimum transition hypothesis(MTH) was implemented. A two-stage hierarchy was adopted to perform the high-level and the low-level con...
详细信息
ISBN:
(纸本)9781538619797;9781538619780
In this study, a hierarchical architecture for the intermittent control under the minimum transition hypothesis(MTH) was implemented. A two-stage hierarchy was adopted to perform the high-level and the low-level control respectively. The high-level controller performed the intermittent control by setting a sequence of goals for the low-level controller. Goal planning as the intermittent control policy was learned with hierarchical deep deterministic policy gradient(h-DDPG) proposed in this study, which is a hierarchical version of the conventional DDPG. The model successfully learned to temporally decompose a complex movement into a sequence of basic motor skills with sparse transitions, as shown in results of the two validation experiments: the trajectory following and the obstacle avoidance tasks.
暂无评论