In this paper, we describe a novel 3D shape retrieval method based on new features. The features are extracted for 3D points based on a 2D attribute space which consists of two bidirectional 3D shape attributes, one o...
详细信息
ISBN:
(纸本)9781479905607
In this paper, we describe a novel 3D shape retrieval method based on new features. The features are extracted for 3D points based on a 2D attribute space which consists of two bidirectional 3D shape attributes, one of which along the shape surface direction and the other along the shape content direction perpendicular to the former. We define this features as Point Bidirectional Features (PBFs), which can reflect not only the global shape but also the local content. For simplicity, we choose the geodesic distance (GD) and the shape diameter function (SDF) to construct PBFs. Given a database of 3D shapes and a query shape, the proposed retrieval method adopts a shape similarity measurement based on 3D points matching with PBFs to decide which is the most similar shape from the database. To reduce cost of computation and feature storage, a shape simplification algorithm is integrated into this method. Additionally, a simple but effective point correspondence mechanism with K-Nearest Neighbor (KNN) assignment is designed for this retrieval method. Experimental results demonstrate the effectiveness of the proposed features and method.
An effective dual-channel noise reduction algorithm is proposed based on sparse representations. The algorithm is composed of the following steps. Firstly, overlapping patches sampled from two channels together instea...
详细信息
An effective dual-channel noise reduction algorithm is proposed based on sparse representations. The algorithm is composed of the following steps. Firstly, overlapping patches sampled from two channels together instead of each channel one by one are trained to be a dictionary via K-SVD. Secondly, OMP(Orthogonal-Matching-Pursuit) reconstruction algorithm is applied to obtain the sparse coefficients of patches using the dictionary. Thirdly, the denoising speech can be obtained by the updated coefficients. Lastly, the above three steps are iterated to get clearer speech until some conditions are reached. Experimental results show that this algorithm performs better than that with single channel.
Recently, extensive research and application on contrast enhancement of radiographs based on multi-scale decomposition of the images have validated its higher performance than regular techniques. However, to some exte...
详细信息
Recently, extensive research and application on contrast enhancement of radiographs based on multi-scale decomposition of the images have validated its higher performance than regular techniques. However, to some extent, conventional multi-scale methods suffered from the introduction of visible artifacts. In this work, we present an algorithm for nonlinear chest radiograph contrast enhancement algorithm within the multi-scale decomposition architecture in spatial domain. In particular, one kind of nonlinear enhancement function is designed by exploiting local contrast information. The main contribution of this model is the local adaptive enhancement ability, which can avoid visible artifacts, while keeping the same detail enhancement ability. In the meantime, no excessive noise is amplified, comparing to conventional methods. Finally, an evaluation using a chest image is provided to demonstrate the effectiveness of the proposed algorithm.
We propose an unsupervised person search method for video surveillance. This method considers both the spatial features of persons within each frame and the temporal relationship of the same person among different fra...
详细信息
We present a novel fast method based on computer vision to identify microbe The proposed method is simple but absolutely effective It combines approximate parallel light source and industrial camera, to automatically ...
详细信息
ISBN:
(纸本)9781479920327
We present a novel fast method based on computer vision to identify microbe The proposed method is simple but absolutely effective It combines approximate parallel light source and industrial camera, to automatically accomplish the bacteria identification and monitor the growing states of bacteria during the progress of a drug sensitive test. Based on this method, the color information and turbidity information, which reflect the primary information of drug sensitive tests, can be obtained fast, while processing efficiency can be as high as hundreds of milliseconds per frame. The performance of our method is significantly accurate and robust.
Neuro-fuzzy(NF)networks are adaptive fuzzy inference systems(FIS)and have been applied to feature selection by some ***,their rule number will grow exponentially as the data dimension *** the other hand,feature select...
详细信息
Neuro-fuzzy(NF)networks are adaptive fuzzy inference systems(FIS)and have been applied to feature selection by some ***,their rule number will grow exponentially as the data dimension *** the other hand,feature selection algorithms with artificial neural networks(ANN)usually require normalization of input data,which will probably change some characteristics of original data that are important for *** overcome the problems mentioned above,this paper combines the fuzzification layer of the neuro-fuzzy system with the multi-layer perceptron(MLP)to form a new artificial neural ***,fuzzification strategy and feature measurement based on membership space are proposed for feature selection. Finally,experiments with both natural and artificial data are carried out to compare with other methods,and the results approve the validity of the algorithm.
Matching of appearance-based object representations using eigenimages is computationally very demanding. Most commonly, to recognize an object in an image, parts of the input image are projected onto the eigenspace an...
详细信息
Tissue texture reflects the spatial distribution of contrasts of image voxel gray levels,i.e.,the tissue heterogeneity,and has been recognized as important biomarkers in various clinical *** computed tomography(CT)is ...
详细信息
Tissue texture reflects the spatial distribution of contrasts of image voxel gray levels,i.e.,the tissue heterogeneity,and has been recognized as important biomarkers in various clinical *** computed tomography(CT)is believed to be able to enrich tissue texture by providing different voxel contrast images using different X-ray ***,this paper aims to address two related issues for clinical usage of spectral CT,especially the photon counting CT(PCCT):(1)texture enhancement by spectral CT image reconstruction,and(2)spectral energy enriched tissue texture for improved lesion *** issue(1),we recently proposed a tissue-specific texture prior in addition to low rank prior for the individual energy-channel low-count image reconstruction problems in PCCT under the Bayesian *** results showed the proposed method outperforms existing methods of total variation(TV),low-rank TV and tensor dictionary learning in terms of not only preserving texture features but also suppressing image *** issue(2),this paper will investigate three models to incorporate the enriched texture by PCCT in accordance with three types of inputs:one is the spectral images,another is the cooccurrence matrices(CMs)extracted from the spectral images,and the third one is the Haralick features(HF)extracted from the *** were performed on simulated photon counting data by introducing attenuationenergy response curve to the traditional CT images from energy integration *** results showed the spectral CT enriched texture model can improve the area under the receiver operating characteristic curve(AUC)score by 7.3%,0.42%and 3.0%for the spectral images,CMs and HFs respectively on the five-energy spectral data over the original single energy data *** CM-and HF-inputs can achieve the best AUC of 0.934 and *** texture themed study shows the insight that incorporating clinical important prior information,e.g.,tiss
Curve matching is one of key issues in computer vision, image analysis and patternrecognition. Based on discrete V-transform, the distance is calculated between curves using the descriptor of V-system to find the mat...
详细信息
Curve matching is one of key issues in computer vision, image analysis and patternrecognition. Based on discrete V-transform, the distance is calculated between curves using the descriptor of V-system to find the matching curves, and then the matching parameters are evaluated in this article. The new approach can find efficiently the rough location of a short extracted image curve in a long reference curve. Different from the existing approaches, it needn't to extract feature points. Extensive tests show that it is efficient.
In this paper, a universal full-reference (FR) image quality metric based on Edge structure similarity (QMESS) is proposed using spatial position displacement degree of wavelet transform modulus maxima between referen...
详细信息
In this paper, a universal full-reference (FR) image quality metric based on Edge structure similarity (QMESS) is proposed using spatial position displacement degree of wavelet transform modulus maxima between reference image and distorted image in multi-resolution domain. Firstly, we decompose images in wavelet domain. The structure error between reference images and distorted images is computed based on the statistics of spatial position error of local modulus maxima in wavelet domain. At the same time, peak signal to noise ratio (PSNR) is adopted to evaluate the stochastic noise in images. Finally, the low frequency resolution layer distortion is evaluated by means of the mutual information and the luminance distortion. The three components are combined for the whole visual distortion measurement. From the experiment results, the proposed metric is much better than conventional PSNR method and the state-of-the-art SSIM approach in terms of the performance relative to subjective judgment. Comparing to the excellent VIF method, the proposed method performs better in individual distortions and obtains similar results on cross-distortion type.
暂无评论