Shuxu Guo

Finger-Vein Recognition Based on the Score Level Moment Invariants Fusion

This paper expresses an algorithm of finger-vein recognition based on the score level moment invariants fusion. With regard the both characteristics of low contrast and intensity inhomogeneity in the infrared vein images; the maximum curvature model is adopted to extract the finger-vein pattern. Then, seven moment invariants are extracted to be matched by the Euclidean distance. In order to solve the problem of high false rate in finger-vein recognition using the single-feature method and the equal weights fusion strategy, the matching scores are fused by the weighted average strategy, and the equal error rate (EER) is minimized to obtain the optimum weights. Finally, the fused matching score is used to make the final decision. Experiment results prove that our algorithm has high performance on recognition rate, which at least reduces 11% in EER compared with the single-feature method and the equal weights fusion strategy.

Automatic segmentation of liver tumors from multiphase contrast-enhanced CT images based on FCNs

This paper presents a novel, fully automatic approach based on a fully convolutional network (FCN) for segmenting liver tumors from CT images. Specifically, we designed a multi-channel fully convolutional network (MC-FCN) to segment liver tumors from multiphase contrast-enhanced CT images. Because each phase of contrast-enhanced data provides distinct information on pathological features, we trained one network for each phase of the CT images and fused their high-layer features together. The proposed approach was validated on CT images taken from two databases: 3Dircadb and JDRD. In the case of 3Dircadb, using the FCN, the mean ratios of the volumetric overlap error (VOE), relative volume difference (RVD), average symmetric surface distance (ASD), root mean square symmetric surface distance (RMSD) and maximum symmetric surface distance (MSSD) were 15.6±4.3%, 5.8±3.5%, 2.0±0.9%, 2.9±1.5mm, 7.1±6.2mm, respectively. For JDRD, using the MC-FCN, the mean ratios of VOE, RVD, ASD, RMSD, and MSSD were 8.1±4.5%, 1.7±1.0%, 1.5±0.7%, 2.0±1.2mm, 5.2±6.4mm, respectively. The test results demonstrate that the MC-FCN model provides greater accuracy and robustness than previous methods.

Vein Pattern Extraction Based on the Position-Gray-Profile Curve

Biometric identification is an important security application and it requires non-intrusive capture and real-time processing. Recently, vein recognition appears to be making real headway in the market, and is considered as the more novel biometric, which is called the Fourth Biometric. This article puts forward a performed solution of a near infrared radiation finger vein imaging processing for vein recognition system. The position-gray-profile (PGP) curve is proposed to analyze the NIR image of vein through the observing. After the stage of the pre- process of the finger vein image, the curvature of the position- gray-profile curve is calculated to extract the vein pattern. In order to test performance of the method, Hausdorff distance is used. Preliminary testing on a database which contains 378 different finger vein images has been carried out and all the images are correctly recognized.

A novel method of dynamic target detection

The shooting range test is an important field in modern weapon development. The modern weaponry is developing towards long distance and automation directions, therefore the shooting range test is put forward new higher requirements. A novel method of target detection based on the digital image processing technology is proposed in the paper. Experiments indicate the strategy is fit to the request of the dynamic target detection and tracking in the shooting range.

Stripe noise removal method for MODIS remote sensing imagery

Stripe noise seriously influences the quality of remote sensing imagery, an effective method for removing stripe noise in MODIS (Moderate Resolution Imaging Spectroradiometer) imagery is proposed in this paper. The proposed method mainly considers the scanning characteristic of multi-detectors in MODIS. Utilizing the high correlation between detector subimages to predict the new detector subimages, then use these new detector subimages to compose the destriped image. Experimental results prove the proposed method is superior to present destriping methods, which can remove stripe noise well and preserve most information of original image. The proposed method is also applicable in stripe noise removal of other multi-detectors remote sensing imagery.

Automatic Organ Segmentation for CT Scans Based on Super-Pixel and Convolutional Neural Networks

Accurate segmentation of specific organ from computed tomography (CT) scans is a basic and crucial task for accurate diagnosis and treatment. To avoid time-consuming manual optimization and to help physicians distinguish diseases, an automatic organ segmentation framework is presented. The framework utilized convolution neural networks (CNN) to classify pixels. To reduce the redundant inputs, the simple linear iterative clustering (SLIC) of super-pixels and the support vector machine (SVM) classifier are introduced. To establish the perfect boundary of organs in one-pixel-level, the pixels need to be classified step-by-step. First, the SLIC is used to cut an image into grids and extract respective digital signatures. Next, the signature is classified by the SVM, and the rough edges are acquired. Finally, a precise boundary is obtained by the CNN, which is based on patches around each pixel-point. The framework is applied to abdominal CT scans of livers and high-resolution computed tomography (HRCT) scans of lungs. The experimental CT scans are derived from two public datasets (Sliver 07 and a Chinese local dataset). Experimental results show that the proposed method can precisely and efficiently detect the organs. This method consumes 38xa0s/slice for liver segmentation. The Dice coefficient of the liver segmentation results reaches to 97.43%. For lung segmentation, the Dice coefficient is 97.93%. This finding demonstrates that the proposed framework is a favorable method for lung segmentation of HRCT scans.

Liver lesion segmentation in CT images with MK-FCN

This paper presented an approach used Fully Convolutional Networks (FCN) to segment liver tumor in Computed Tomography (CT) images. In addition, using different characteristics of scan quality and tumor conspicuity among portal venous phase, arterial phase and equilibrium phase, we proposed an automatic liver tumor segmentation with Multiple Kernel Fully Convolutional Networks (MK-FCN). MK-FCN can segment liver tumor from multi-phase contrast-enhanced CT images by using different characteristics of scan quality and tumor conspicuity among different phases. Experiments proved the effectiveness of this method in the liver tumor segmentation.

Shape feature extraction using dual-tree complex wavelet moment invariants method

In this paper, we proposed a novel method to extract shape feature based on dual-tree complex wavelet. First, with the two level dual-tree complex wavelet transformations, we can get two low frequency components of the first level, which are used as wavelet moment invariants formed from approximation coefficients. Then, we calculate means and variance for each of the six detailed components in the second level since it contains different directions information of the shape. Using the Principal Component Analysis (PCA), twenty features can be reduced to five maximum useful features which contribute to shape matching.

Finger vein image deblurring reconstruction based on Polak-Ribière conjugate Gradient Projection

In this paper, we presents a compressive sensing (CS) based application --Polak-Ribière conjugate gradient projection (PR-CGP) for solving bound constrained quadratic program to reduce noise in synthetic vein images and real finger vein images respectively which blurred with various noise. Then compares the result with the reconstruction by Gradient Projection for Sparse Restruction(GPSR) algorithm. The results show that the method has better performance than the GPSR method in reducing noise, and thus provides more accurate information for vein recognition and extraction.

Finger Vein Image Denoising Based on Compressive Sensing

To extract venous information from noise-added images acquired by infrared sensor, in this paper, we presents a compressive sensing (CS) based application -- gradient projection for sparse reconstruction (GPSR) -- to reduce noise in synthetic vein images and real finger vein images respectively. Then compares the result with the reconstruction by wavelet threshold denoising algorithm. The results show that the GPSR has better performance than the wavelet threshold method in reducing noise, avoids losing the edge information of finger vein, and thus provides more accurate information for vein recognition and extraction.