In this paper, we propose an AdaBoost Random Forest (AdaBoost-RF) based blind multiply distorted image quality assessment (IQA) method. The AdaBoost-RF is an ensemble learning algorithm that uses the principle of the ...
详细信息
ISBN:
(纸本)9781538619377
In this paper, we propose an AdaBoost Random Forest (AdaBoost-RF) based blind multiply distorted image quality assessment (IQA) method. The AdaBoost-RF is an ensemble learning algorithm that uses the principle of the AdaBoost and random forest as Weak Learners. We also operate learning quality-aware features (LQAF) and utilize the AdaBoost-RF to get the image quality score by taking advantage of image features. The AdaBoost-RF increases the degree of difference between Weak Learners, in contrast to other regressions, and can gain higher predictive accuracy. On the LIVE multiply distorted image database (LIVEMD) and MDID2013, experimental results show our proposed model has a better performance than the other mainstream IQA methods, and is a useful and reliable method for image quality assessment.
The optical flow field-based active demons algorithm is a major technique for nonrigid image registration. Current active demons algorithms for large deformation problems are suffered with low registration precision d...
详细信息
ISBN:
(纸本)9781538619377
The optical flow field-based active demons algorithm is a major technique for nonrigid image registration. Current active demons algorithms for large deformation problems are suffered with low registration precision due to using gradient information of the reference image and moving image only. To avoid the disadvantage of the active demons algorithms, this paper first presents the length and area of connected region as driving forces into the active demons diffusion equation and then proposes a new deformable image registration algorithm by combining the gradient information-based driving forces and connected region-based driving forces. Furthermore, a multi-resolution technique is introduced for robust performance against noise. Experimental results show that the proposed deformable image registration algorithm has better performance than several conventional active demons algorithms.
Stereo matching plays an important role in many applications, such as Advanced Driver Assistance Systems, 3D reconstruction, navigation, etc. However it is still an open problem with many difficult. Most difficult are...
详细信息
ISBN:
(纸本)9781538619377
Stereo matching plays an important role in many applications, such as Advanced Driver Assistance Systems, 3D reconstruction, navigation, etc. However it is still an open problem with many difficult. Most difficult are often occlusions, object boundaries, and low or repetitive textures. In this paper, we propose a method for processing the stereo matching problem. We propose an efficient convolutional neural network to measure how likely the two patches matched or not and use the similarity as their stereo matching cost. Then the cost is refined by stereo methods, such as semiglobal maching, subpixel interpolation, median filter, etc. Our architecture uses large image patches which makes the results more robust to texture-less or repetitive textures areas. We experiment our approach on the KITTI2015 dataset which obtain an error rate of 4.42% and only needs 0.8 second for each image pairs.
To solve the insufficient of nonlinearity in optical cryptosystems, a new scheme is proposed for optical image encryption based on Quantum Cellular Neural Network(QCNN) hyperchaotic system and Anamorphic Fractional Fo...
详细信息
ISBN:
(纸本)9781538619377
To solve the insufficient of nonlinearity in optical cryptosystems, a new scheme is proposed for optical image encryption based on Quantum Cellular Neural Network(QCNN) hyperchaotic system and Anamorphic Fractional Fourier Transform(AFrFT). The first and the second Chaos Random Phase Masks(CRPM) are generated by QCNN hyperchaotic system. The original image is combined with the first CRPM and then transformed by AFrFT. The encrypted image is obtained by the second CRPM and the second AFrFT. Experimental simulation results show the feasibility, high security and efficiency of our scheme.
The traditional method of mine belt conveyor fire accident detection is limited. A novel imageprocessing method is proposed in this paper, which combines 2D-Dimensional Otsu, Canny dge detection and Artificial Bee Co...
详细信息
ISBN:
(纸本)9781538619377
The traditional method of mine belt conveyor fire accident detection is limited. A novel imageprocessing method is proposed in this paper, which combines 2D-Dimensional Otsu, Canny dge detection and Artificial Bee Colony algorithm. The noise interference of image is reduced by using median filtering techniques. In order to obtain gray value and neighborhood gray scale average, 2-Dimensional histogram of image is constructed. The fitness function of Artificial Bee Colony algorithm is designed by 2-Dimensional Otsu method and the maximum value of the fitness function is the optimal threshold for image segmentation. Finally, the canny edge detection and grayscale morphology were used to extract the target. Theoretical analysis and simulation results show that the proposed method is effective to detect fire accident of mine belt conveyor in complex undeground environment.
Many evolution behaviors can be described by nonlinear partial differential equations. PDE-based models have some advantages over many traditional models in complex imageprocessing while considerable computation is o...
详细信息
ISBN:
(纸本)9781538619377
Many evolution behaviors can be described by nonlinear partial differential equations. PDE-based models have some advantages over many traditional models in complex imageprocessing while considerable computation is often required. This paper presents a new model based on linear diffusion equation and noise estimation. The local diffusion coefficient can be well determined by the estimation of local standard variation of the noise. The denoising results can be visibly improved to approximate even exceed some results of nonlinear models(PM, TV, etc.) without amounts of computation. Some numerical experiments show the accuracy and efficiency of proposed model.
Matching image feature point by the SURF (Speed-up Robust Features) algorithm needs to loop through all feature points on the image to be matched, which will take a long computation time. In order to solve this proble...
详细信息
ISBN:
(纸本)9781538619377
Matching image feature point by the SURF (Speed-up Robust Features) algorithm needs to loop through all feature points on the image to be matched, which will take a long computation time. In order to solve this problem, it is proved based on the photo consistency that the Hessian matrix determinants between the matched feature points are equal in theory. Experimental results show that because of the influences of light and other elements, the ratio of the Hessian matrix determinant of 95 percent of the correct SURF feature point's pairs is between 0.7 and 1.5. Based on the relation of the Hessian matrix determinants between the matched feature points, a fast image feature point matching algorithm is proposed, which improves the recognition rate of feature points, and it makes over twice faster than the SURF algorithm.
The multispectral remote sensing image sensor is expensive, and it will cause part of the pixels of the defect in the process of acquisition and transmission, it would be of great importance to be able to restore the ...
详细信息
ISBN:
(纸本)9781538619377
The multispectral remote sensing image sensor is expensive, and it will cause part of the pixels of the defect in the process of acquisition and transmission, it would be of great importance to be able to restore the defective pixels at the receiver. According to the latest research results of optimization, this paper presents a new multi-spectral remote sensing image restoration method based on sparse representation. The method can divide three-dimensional image into different blocks and model the problem of multi-spectral remote sensing image, and the multi-spectral pixel blocks of the study area is restored by sparse approximation. The experiment proves the efficiency of the algorithm, and the proposed method is very important in remote sensing imageprocessing.
In this paper, we address the problem of image reconstruction from highly undersampled Fourier measurements. In order to promote inherent sparsity in gradient and wavelet transform domain, we proposed a new reconstruc...
详细信息
ISBN:
(纸本)9781538619377
In this paper, we address the problem of image reconstruction from highly undersampled Fourier measurements. In order to promote inherent sparsity in gradient and wavelet transform domain, we proposed a new reconstruction scheme via minimizing L-0 norm of gradients and L-1 norm of wavelet coefficients. L-0 gradient minimization can control the number of non-zero gradients to enforce the sparsity in gradient, which results in edge preserving reconstruction. The reconstruction is casted into optimization framework and alternating direction method of multipliers (ADMM) algorithm is utilized to efficiently solve the proposed optimization problem. Experimental results demonstrate the superior performance of the proposed method in comparison with the L-1 gradient reconstruction method.
In order to realize the configuration convenience and management standardization of the image gateway, this paper adopts a scheme of configuration and management by using the embedded system and the host computer soft...
详细信息
ISBN:
(纸本)9781538619377
In order to realize the configuration convenience and management standardization of the image gateway, this paper adopts a scheme of configuration and management by using the embedded system and the host computer software. The scheme uses K60 (NXP Freescale MK60DN512VLQ10) as the master chip, transplants RT-Thread, a real-time operating system, and a lightweight protocol stack LwIP, which runs TCP / UDP Socket service, listens to the host computer command. The embedded system is connected with the core processor FPGA (Field Programmable Gate Array) through the RMII and the UART interfaces. To realize the configuration of the image gateway, the embedded system reads the network packets from FPGA through the RMII interface and sends the security rules to FPGA through the UART.
暂无评论