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.
Aiming at the problem that traditional Meanshift tracking algorithm cannot track target accurately. When target is occluded or background interferences appear, a new tracking algorithm based on features matching and motion detection predicting is proposed. The new algorithm extracts features of target by the SIFT algorithm to realize the location. The original iteration position is estimated by the...
The present status of heart sound recognition is introduced in the paper. In order to improve the performance of heart sound recognition, a new model based on SVM is proposed. Firstly, the wavelet transform is used to reduce the noise of the heart sound, and then MFCC feature is extracted from heart sound. On this basis, the Support Vector Machine is used to build the classification model. In the...
The purpose is to locate the fault of coal mine drainage system quickly and accurately, and then establish a fault diagnosis of the underground drainage fuzzy Petri net system model based on BP optimization algorithm. On the basis of the traditional fuzzy Petri net model, the BP optimization algorithm with momentum factor is added to improve the accuracy of the algorithm, the S-type continuous functions...
In order to remove the false edge points extracted by the Canny operator in etched character recognition, a method based on the simplified neighborhood feature and AdaBoost algorithm was proposed. In conventional neighborhood feature, canny edge points are taken as the center and neighborhood pixels are extracted as the feature. However, the dimension of the neighborhood feature rises with the square...
In the field of video-based fire detection, traditional fire color models have poor adaptability and susceptibleness to environmental interference. As fire can be divided into fire core, inner fire and outer fire layers depending on its temperature and color, a new fire detection model based on dispersion of fire color components has been proposed by the author. In order to further reduce false alarm,...
Network analysis of signals originating from different parts of brain during motor imagery (MI) has gained lots of interest recently. In this paper, we used EEG to construct the brain network during MI, and analyzed the topological characteristics of the EEG function network. It is found that the node degree and clustering coefficient of the right hand MI is higher than the left hand MI on the nodes...
Problems related to fault detection of turnouts are discussed in this paper, which is important to guarantee the security of a driving train. Motivated by an existing approach for fault detection of turnouts and to deal with its demerits, a hybrid feature extraction method is given, which extracts information in both time and frequency domains, then support vector machine (SVM) is further used to...
Sensor plays an important role in complex industrial environment. Therefore, researches on sensor fault diagnosis technology are important for improving the reliability of industry system. A sensor signal which is non-linear and non-stationary, has many kinds of structural characteristics and sensing properties. On the basis of supervised locally linear embedding (SLLE), support vector machine (SVM),...
In view of the emotional polarity classification problem, the deep learning has the disadvantages of incomplete information extraction and low precision, a model combining bi-directional gated recurrent unit with multiple convolution neural network is proposed. The unit is used to extract the history and future information of the sentence, then use the multi-convolution neural network for system training,...
Low-resolution (LR) is a challenging problem in the real world. In order to obtain better performance for low-resolution face recognition (LRFR), this paper employs a novel approach for matching low-resolution images with high resolution (HR) images based on two-dimensional linear discriminant analysis (2D-LDA) and metric learning method. The LR and HR images are transformed into a common space via...
The interoperability testing of CTCS-3 Level Train Control System guarantees the safe operation of train running on different lines. It makes great sense to achieve automatic analysis of interoperability testing results, which could improve the efficiency and accuracy of testing. In this paper, a research was conducted on automatic analysis of testing results for on-board equipment of train control...
Moving object tracking with discriminative model is very popular in recent years, which focuses on online selecting highly informative features to maximize the separability between object and background. An adapted particle filter tracker with online learning and inheriting discriminative model is proposed in this paper. Top-ranked discriminative features are selected into appearance model by Online...
Robotic simultaneous localization and mapping (SLAM) confronts extreme challenge in collapsed, cluttered, GPS-signal unreliable environments of search and rescue (SAR). Our improved SLAM methods aim to mobile robot performing SAR requirements which comprise the significant objects identification, loop closure perceiving, exploration area coverage, and the other performances. We developed efficient...
Despite great progress has been made in recent years, efficient and robust people detection continues to be a challenging problem in the filed of computer vision. In this paper, we propose a highly efficient indoor people detect method based on RGB-D sensor. First, two RGB and depth feature fusing strategies are proposed and compared. Secondly, an improved non-maximum suppression algorithm is proposed...
Traditional vehicle detectors always utilize singletemplate model to represent the vehicle which can not encircle vehicles with different aspect ratios. In this paper, we propose a fast and accurate approach for detecting vehicles which joints classification and aspect ratio regression. The key idea is extending the boosting decision trees method to estimate vehicle's aspect ratio during vehicle detection,...
Action prediction is an important item in the fields of pattem recognition and computer vision. Capturing evolution tendency cues is a key point to action prediction. Dense trajectory (DT) and dynamic image are both effective approaches to explore dynamic information in videos. DT is used to describe the features of action and has achieved state-of-the-art results on action recognition, but often...
Road markings are important information of transport systems for drivers or intelligent vehicle. Efficient road markings feature extraction is pre-requisite to road markings detection, recognition and visual localization. However, most of previous lane markings feature extractors are operating on conventional images, the feature extraction methods for omnidirectional images are rarely considered in...
As a special task beyond general object detection, pedestrian detection has attracted much attention in recent years. Despite the significant improvements, detecting pedestrians is still a challenging task, especially for small-size pedestrians. In this paper, we present a multi-scale feature fusion convolutional neural network (MFF-CNN) for pedestrian detection. The MFF-CNN is benchmarked on three...
Crowd counting on still images is very challenging due to heavy occlusions and scale variations. In this paper, we aim to develop a method that can accurately estimate the crowd count from a still image. Recently, convolutional neural networks have been shown effective in many computer vision tasks including crowd counting. To this end, we propose a fully convolutional network (FCN) architecture to...
Flow pattern is one of the most important parameters for gas-liquid two-phase flow. In this work, a new flow pattern identification method based on Convolution Neural Network (CNN) is presented. A 7-layer CNN structure is chosen, and the parameters of this network are determined by a training set. In order to verify the feasibility, experiments were carried out in horizontal pipe with the inner diameter...
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.