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.
In this paper, a new method was introduced in the Chinese license plate recognition. We propose a convolutional neural network architecture designed to recognize license plate directly from pixel images with no preprocessing. We present the image transformation applied on the original license plate to increase the training database. We also provide experimental results to demonstrate the robustness...
Two-dimensional principal component analysis technique is an important and well-developed area of image recognition and to date this method has been put forward. A new face recognition method two-dimensional principal component analysis (2DPCA) based on BP neural networks, named 2DPCA-BP method, was proposed. 2DPCA was used to obtain a family of projected feature vectors, in which face image was projected...
The thresholding method is a fundamental part for motion detection and other advanced applications,so this step is very important, it must base on a reliable, effective method in order to access to the robustness of the computer intelligent video surveillance system. A new thresholding method which is simple and effective is proposed in the article. Firstly, we propose an imitating uni-Gaussian model...
Wall rock alteration is one of the major mineralization characteristics of hydrothermal deposits. In order to effectively extract information in hyperspectral remote sensing prospecting, it is necessary to objectively filter the spectral characteristics of altered rock and the closely correlative spectral characteristics in the mineralization forecast. Rough sets do not need the datapsilas additional...
A face recognition method based on fuzzy data fusion is presented. In traditional principle component analysis method, operating directly on the whole face image leads to only global information about face image can be extracted and local one may be neglected. It is not very effective under variations of facial expression, pose and illumination. To solve this problem, in proposed scheme, each original...
A new approach is proposed to improve face recognition in the paper. The accurate face is detected by the position relation of the face and the eyes. The face features are extracted using the Gabor wavelet and the Adaboost algorithm is used to detect the face and the eyes. In the actual detection of the face, the face is probably inclining, and then we correct the detected face according to the positions...
On the basis of wavelet and singular signal characteristic analysis, a new defect detection method based on wavelet characteristics is presented. Firstly, to select wavelet filters of compact support and high vanishing moment properties are regarded as finite biorthogonal filters. Secondly, the new wavelet with concussive and filters coefficients of centralized distribution is constructed by lifting...
In this paper, a robust and effective face detection method with HTF-Boosting is proposed. Firstly, a new feature, called Haar texture feature, is proposed that has many merits compared with Haar-Like feature. Secondly, a new Boosting algorithm, called Haar Texture Feature Boosting (HTF-Boosting), is proposed to construct strong face/nonface classifiers. The HTF-Boosting algorithm trains strong classifiers...
A feature fusion algorithm with application to facial expression recognition is presented. Firstly, the brows, eyes and mouth areas are segmented from the facial expression images, and are computed with Higher-order Local Auto-Correlation (HLAC) method, and the Weighted Principal Component Analysis (WPCA) is used to reduce dimensions secondly, in which the weights values are obtained according to...
The basic principle of online Chinese signature identification is to compare one's signature characteristic value with those in the appointed pattern librarypsilas so that the signature to be identified is classified as either authentic or forged. In order to form a stable and effective pattern library, a new method is proposed based on the fact that the Fisher distinction criterion function is a...
A digital watermarking method for color image in wavelet domain is proposed. The watermarking information which include a section of template information, is embedded into low-frequency coefficients of Y component in YUV color space. This section of template information defines a kind of relation between the low-frequency coefficients corresponding to the four sub-images obtained by down-sampling...
Combined with audio features and the advantage of discrete multiwavelet transform (DMT) in signal processing, a novel audio watermarking scheme is proposed in this paper. Due to the high relevance between adjacent audio frames, the watermark embedding and detection are based on the energy relationship of two consecutive frames in multiwavelet domain. To enhance the robustness and auditory quality,...
In this paper, we proposed a novel approach to image completion for overlapping chromosomes. In our system, only given missing regions, the task can be performed automatically without human intervention. We address the problem of image completion for overlapping chromosomes in the context of a discrete global optimization problem. In order to reconstruct the original chromosome image as faithfully...
It is known to all that obtaining an effectual feature representation is of paramount importance to face recognition. In this paper, the latest feature extraction method based on KCCA is introduced. However, in the training stage of the standard KCCA-based extractor, it requires to store and manipulate the kernel matrix, the size of which is square of the number of samples. When the sample numbers...
This paper describes a modified GMM based voice conversion method,in order to avoid the over-smoothed phenomena of traditional GMM conversion method, we adopt GMM method considering posterior threshold of each frame. We also considered the dynamic spectral features between frames to overcome the speech discontinuity problem,experiments show that our presented GMM method has shown better performance...
In this paper, a novel method is proposed to detect faces based on PCNN time signature and skin color segmentation, in which no training is needed. A test image is first divided into overlapped blocks and extracted PCNN time signature as the detection features, which a two-dimensional image is projected to a one-dimensional feature space. The test blocks are matched to a face template, which can be...
This paper presents a novel method for fingerprint image quality. Five features are extracted from the fingerprint image to analyze the quality and the feature vector is formed from the five features. Then SVM classifier, which can solve small-sample learning problems with good generalization, is trained to classify the fingerprint image. The fingerprint image is separated into one of the three classes,...
An eye-gaze tracking system-EyeSecret for real scene, is developed based on pupil-cornea reflection vector method. The general structure and software design are introduced. In the hardware of EyeSecret, design and optimization of optical system and control circuits were proposed. In the software, implementation algorithms of key modules, cornea reflection detection, pupil center extraction and calibration,...
In this paper, we present a novel scene change detection algorithm in the H.264/AVC compressed domain directly. Three major statistical features are used: (i) bit allocation for intra mode macroblocks (BIM), (ii) bit allocation for inter mode macroblocks (BPM), and (iii) the number of skip mode macroblocks in a frame. These features can be easily extracted from the H.264/AVC bitstreams. Besides the...
This paper presents a novel lane detection algorithm for automatic drive system. The algorithm chooses a common curved lane parameter model which can describe both straight and curved lanes. The most prominent contribution of this paper is: instead of using one single method to calculate all the parameters in the lane model, both the adaptive random Hough transformation (ARHT) and the Tabu Search...
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.