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 the process of aircraft residual ice detection based on near infrared spectrum method, the qualities of the images, such as brightness and contrast, are easily influenced by sunlight, weather and other factors. Therefore, it is hard to precisely match the feature points. In this paper, an improved Harris local invariant feature matching method is proposed. The improved Harris feature point detection...
We present a new Cascaded Shape Regression (CSR) architecture, namely Dynamic Attention-Controlled CSR (DAC-CSR), for robust facial landmark detection on unconstrained faces. Our DAC-CSR divides facial landmark detection into three cascaded sub-tasks: face bounding box refinement, general CSR and attention-controlled CSR. The first two stages refine initial face bounding boxes and output intermediate...
Different kinds of features hold some distinct merits, making them complementary to each other. Inspired by this idea an index level multiple feature fusion scheme via similarity matrix pooling is proposed in this paper. We first compute the similarity matrix of each index, and then a novel scheme is used to pool on these similarity matrices for updating the original indices. Compared with the existing...
To avoid the introduction of false information during the fusion progress, a novel multi-focus image fusion method is proposed in quaternion wavelet transform domain. To obtain the dependency in different high frequency subbands, a quaternion wavelet contextual hidden Markov model (Q-CHMM) is established for modeling quaternion wavelet coefficients. And for better image representations, several features...
Contrast change is a special type of image distortion which is vitally important for visual perception of image quality, while little investigates has been dedicated to the contrast-distorted images. A proper contrast change not only reduces human visual perception, instead of improving it. This characteristic determines that full-reference way cannot assess contrast-distorted images properly. In...
Convolutional neural networks(CNN) has achieved fairly good results in the field of computer vision. In recent years, with the rapid development of deep learning, more and more researchers tried to apply CNN to the field of Natural Language Processing(NLP). This paper uses CNN model to analyze the sentiment of Chinese text, and improves the structure of basic CNN, adjusts different parameters to carry...
An information cascade occurs when a person observes the actions of others and then engages in the same acts. Cascades may break out if a large population of nodes in the network get affected. The outbreaks of cascades will often bring influential events, which leads to an open research problem: how to accurately predict the cascading outbreaks in social networks? Although there have been some existing...
In this paper, a novel image descriptor, called Color Binary Correlation (CBC), is proposed for image retrieval. This method defines and describes the structure elements utilizing binary patterns based on colors and edge orientation respectively, and thus integrate texture with the other two properties. Besides, its variants CBCri and CBCu2, which are presented for rotated invariance and “unform”...
This paper introduces a regularization method called Correlative Filter (CF) for Convolutional Neural Network (CNN), which takes advantage of the relevance between the convolutional kernels belonging to the same convolutional layer. During the process of training with the proposed CF method, several pairs of filters are designed in a manner of randomness to contain opposite weights in low-level layers...
Visually perceiving human motion at semantic level is an important however challenging problem in multimedia area. In this work, we propose a novel approach to map the low-level responses from visual detection to semantically sensitive description to human actions. The feature map is triggered by the output of deformable part model detection, in which the critical information about body parts configuration...
A textural feature extraction algorithm was proposed to automatically find candidate objects in the selected volume of interest (VOI) and compute textural features on multiplanar images for classification of the mesh and fascia. Firstly, candidate objects were found out in axial plane (A-plane) and coronal plane (C-plane) images with the preprocessing stage. Secondly, textural features of candidate...
This paper presents a new relevance feedback scheme, which incorporates Extreme Learning Machine (ELM) to content-based image retrieval (CBIR) with relevance feedback. Relevance feedback schemes based on Support Vector Machine (SVM) have been proposed in previous paper. However, the performance of the schemes are often poor which is caused by the low speed of SVM algorithm in high dimension data....
This paper presents a novel framework for Content Based Image Retrieval(CBIR), which combines color, texture and spatial structure of image. The proposed method uses color, texture and spatial structure descriptors to form a feature vector. Images are segmented into regions to extract local color, texture and CENTRIST(CENsus Transform hISTogram) features respectively. Multiple-instance learning (MIL)...
A novel image fusion algorithm performed on the feature level is proposed incorporating with region segmentation and Cauchy convolution. Firstly, the fuzzy c-means clustering algorithm(FCM) is used to segment the image in the space of feature difference, which is formed by dual-tree discrete wavelet transform(DT-DWT) sub-bands. Secondly, the high frequency coefficients are modeled by the convolution...
In this paper, two new methods: ECA and 2DECA are proposed for face recognition, which are inspired by KECA. In ECA (2DECA), features are selected in PCA (2DPCA) subspace based on the Renyi entropy contribution instead of cumulative variance contribution. Then the proposed methods are tested on the OLR, YALE and XM2VTS databases respectively. We also compare the performance of the related methods...
Precise fundus image features detection is an important factor for screening diabetic retinopathy. Some noises in fundus image features extraction need to be solved. This paper studies using phase information to attempt get better effect. We use and compare four phase-based approaches and get some instructive results.
For License Plate Recognition (LPR) system, license plate character recognition rate is seriously affected by image quality. To resolve this problem, a new algorithm is proposed, in which Intersecting Cortical Model (ICM) is applied into license plate character recognition. ICM was derived from several visual cortex models, which can be applied to image feature extraction efficiently. In the new algorithm...
A new video surveillance object recognition algorithm is presented, in which improved invariant moments and length-width ratio of object are extracted as shape feature, while color histograms of object are utilized as color feature. On the combination of shape and color features, object recognition is achieved. Based on the algorithm, an intelligent video surveillance system is implemented. Test results...
This paper describes one valid framework of integrating intensity discontinuity and edge continuity for image Straight-line Extraction. First, multiple "edge contours" are traced out from digital image by means of classic LOG algorithm and contour tracing principle. In this step, Zero-Cross points are detected without implementing traditional "edge labeling" procedure and thus,...
Low-cost high-solution multispectral imagery is attractive to various users. This paper presents our ongoing research on developing Unmanned Airship on Board High-resolution multispectral Imaging System for quick-response to time-critical applications. First, the core component of Multispectral Imaging system is described. After that, automatic registration of collected multispectral imageries is...
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.