The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Training deep neural networks is difficult for the pathological curvature problem. Re-parameterization is an effective way to relieve the problem by learning the curvature approximately or constraining the solutions of weights with good properties for optimization. This paper proposes to reparameterize the input weight of each neuron in deep neural networks by normalizing it with zero-mean and unit-norm,...
Although many methods are available to forecast short-term electricity load based on small scale data sets, they may not be able to accommodate large data sets as electricity load data becomes bigger and more complex in recent years. In this paper, a novel machine learning model combining convolutional neural network with K-means clustering is proposed for short-term load forecasting with improved...
A variant of adaptive worst-case (WC) beamformer is devised in this paper, which is robust against arbitrary unknown signal steering vector (SSV) mismatches. Compared with the conventional WC beamforming approach, the proposed method is further improved in terms of robustness by reconstructing the interference-plus-noise covariance matrix (IN-CM) and adaptively adjusting the uncertainty level of the...
Most existing vision-based methods for gaze tracking need a tedious calibration process. In this process, subjects are required to fixate on a specific point or several specific points in space. However, it is hard to cooperate, especially for children and human infants. In this paper, a new calibration-free gaze tracking system and method is presented for automatic measurement of visual acuity in...
Semi-supervised learning and active learning are important techniques to build more accurate model while labeled data are scarce. The objective of this paper is combining both to effectively relieve user labor for multi-class annotation. We propose a novel graph-based active semi-supervised learning framework which aim at efficiently learning a multi-class model with minimal human labor. In particular,...
A novel learning based framework for efficient heterogeneous faces synthesis is proposed. Based on the same spectral distribution of each modality, a statistical probability model is developed for the mapping learning problem between two groups of facial appearances, instead of the traditional linear regression model. Furthermore, in order to eliminate the influences of facial structure and spectrum...
Synthetic aperture radar (SAR) can be used to distinguish areas of contrasting backscatter on glaciers and relate these areas to glacier facies. In the ablation season, there are two typical facies on temperate mountain glaciers: wet snow and ice. The boundary of wet snow and ice is defined as the transient snow line (TSL), which is an important concept in glaciology. In this letter, a new TSL detection...
Due to market demands, universities have increasingly been focusing on students' practice ability raise, to economic management professional students, they always lack of a complete school training system, the mode of which is low cost, relatively simple to operate, and achieved better results easily. Cluster technology is widely used in computer field, with the cluster concept for constructing school...
Crowd density analysis is crucial for crowd monitoring and management. This paper proposes a novel method for crowd density analysis. According to the framework, input images are firstly divided into patches, and each patch is associated with a density label based on its texture features. Finally, local information is synthesized for global density estimation. Local image content is described by features...
Many previous image processing methods discard low-frequency components of images to extract illumination invariant for face recognition. However, this method may cause distortion of processed images and perform poorly under normal lighting. In this paper, a new method is proposed to deal with illumination problem in face recognition. Firstly, we define a score to denote a relative difference of the...
Drawing upon previous literatures in training and downsizing issues, this paper aims at examining how the efficiencies of training programs in downsizing environments might be determined. The authors looked at both pre-training and post-training phase and identified pre-training motivation and transfer of training as two critical factors deciding training efficiencies. Variables such as career planning,...
This paper discusses risk assessment of supply chain based on BP neural network. The risk assessment procedure is discussed and after the risk factors of supply chain identification and analysis, the risk assessment model is built with BP neural network. Through training of the model using MATLAB neural network toolbox and testing the model shows the preciseness and comprehensive practicability.
This paper researches on the issue of computer recognition to the handwritten character images, including lowercase letters and Arabic numerals. In this paper, we preprocess on characters in order to unified the basic features. And then, we apply the basic method of making the grids to extract the features of character, and classify the respectives. At last, we apply the latest heuristic modifications...
OPC is difficult to identify the released bubble in the process of the sample flowing, thus affecting the counting efficiency. This paper presents a bubble Identification method based on ANN, by increasing the number of detection angle and choosing the lateral light structure, conducive to ANN processing FD samples, according to the different refractivity to distinguish bubble. Simulation results...
In this paper, a new feature for text verification is proposed. The difficulties for the selection of features for text verification (FTV) are first discussed, followed by two principles for the FTV: the FTV should minimize the influence of backgrounds, and it should also be expressive enough for all the texts varied in structures prominently. In this paper, we exploit different block partition methods...
Automatic target recognition (ATR) is an important task in image application. A classifier for the airplane recognition based on the merits of rough set theory (RST)and directed acyclic graph support vector machines (DAGSVM) is proposed in this paper. RST can mine useful information from a large number of data and generate decision rules without prior knowledge. DAGSVM have better classification performances...
Crowd estimation is crucial for crowd monitoring and control. It differs from pedestrian detection or people counting in that no individual pedestrian can be properly segmented in the image. This paper describes a novel and efficient system for crowd density estimation, based on local image texture analysis. A novel indication of local binary pattern feature vector called Advanced LBP is proposed...
Local binary pattern (LBP) is a powerful texture descriptor that is gray-scale and rotation invariant. In this paper, an extension of the original LBP is proposed. LBP operator is adopted in multi-layer block domain, instead of pixel domain. Meanwhile, feature dimension is effectively reduced by dual-histogram LBP (DH-LBP). Combining merits of the two, we propose the advanced LBP (ALBP) and use that...
In this paper, a novel framework for face recognition based on discriminatively trained orthogonal rank-one tensor projections (ORO) and local binary pattern (LBP) is proposed. LBP is an efficient method for extracting shape and texture information and it is robustness to illumination and expression, while ORO has been successful in appearance based face recognition by finding orthogonal tensors....
This paper proposes an extension of marginal fisher analysis (EMFA) for dimensionality reduction and analyzes some properties of both EMFA and linear discriminant analysis (LDA), and finally suggests a synthesized discriminant projection (SDP). SDP takes both global class relationship and local geometry structure into account, which maximizes the distance between marginal points and the distance between...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.