Feature aggregation is a crucial step in many methods of image classification, like the Bag-of-Words (BoW) model or the Convolutional Neural Networks (CNN). In this aggregation step, usually known as spatial pooling, ...
详细信息
ISBN:
(纸本)9781509060344
Feature aggregation is a crucial step in many methods of image classification, like the Bag-of-Words (BoW) model or the Convolutional Neural Networks (CNN). In this aggregation step, usually known as spatial pooling, the descriptors of neighbouring elements within a region of the image are combined into a local or a global feature vector. The combined vector must contain relevant information, while removing irrelevant and confusing details. Maximum and average are the most common aggregation functions used in the pooling step. To improve the aggregation of relevant information without degrading their discriminative power for classification in this work we propose the use of Ordered Weighted operators. We provide an extensive evaluation that shows that the final result of the classification using OWA aggregation is always better than average pooling and better than maximum pooling when dealing with small dictionary sizes.
Currently decision-making systems get widespread. These systems are based on the analysis video sequences and also additional data. They are volume, change size, the behavior of one or a group of objects, temperature ...
详细信息
ISBN:
(数字)9781510608986
ISBN:
(纸本)9781510608979;9781510608986
Currently decision-making systems get widespread. These systems are based on the analysis video sequences and also additional data. They are volume, change size, the behavior of one or a group of objects, temperature gradient, the presence of local areas with strong differences, and others. Security and control system are main areas of application. A noise on the images strongly influences the subsequent processing and decision making. This paper considers the problem of primary signal processing for solving the tasks of image denoising and deblurring of multispectral data. The additional information from multispectral channels can improve the efficiency of object classification. In this paper we use method of combining information about the objects obtained by the cameras in different frequency bands. We apply method based on simultaneous minimization L2 and the first order square difference sequence of estimates to denoising and restoring the blur on the edges. In case of loss of the information will be applied an approach based on the interpolation of data taken from the analysis of objects located in other areas and information obtained from multispectral camera. The effectiveness of the proposed approach is shown in a set of test images.
In object identification image Segmentation is the first step in digital imageprocessing. It can be used to compress different segments or areas of image. A novel sub-Markov Random Walk (subRW) algorithm with label p...
详细信息
Smart video surveillance of indoor and outdoor scenes is a challenging task for modern surveillance systems. Different imaging conditions like bad illumination, adverse weather, etc., makes the surveillance process di...
详细信息
ISBN:
(纸本)9781538618950
Smart video surveillance of indoor and outdoor scenes is a challenging task for modern surveillance systems. Different imaging conditions like bad illumination, adverse weather, etc., makes the surveillance process difficult. Recently, researchers have proposed smart surveillance systems with additional features for more accurate monitoring of events, but not much attention is paid to improve the system such that the monitoring process consumes as minimum resources as possible. In this paper, we propose a novel surveillance system that enhances visibility in adverse weather conditions and summarizes the captured videos automatically to reduce storage space. As the summarization process is based on the events in a scene, video interpretation becomes fast and easy. We propose perceptual features that can be used for more meaningful and robust summarization of the video than the existing summarization algorithms. We test the system for both indoor and outdoor scenes and show that the system works well even with multiple moving objects and complex motions.
Fast and accurate digital computation of the fractional Fourier transform (FRT) and linear canonical transforms (LCT) are of utmost importance in order to deploy them in real world applications and systems. The algori...
详细信息
ISBN:
(纸本)9781538615423
Fast and accurate digital computation of the fractional Fourier transform (FRT) and linear canonical transforms (LCT) are of utmost importance in order to deploy them in real world applications and systems. The algorithms in O(NlogN) to obtain the samples of the transform from the samples of the input function are presented for several different types of FRTs and ICTs, both in 1D and 2D forms. To apply them in imageprocessing we consider the problem of obtaining sparse transform domains for images. Sparse recovery tries to reconstruct images that are sparse in a linear transform domain, from an underdetermined measurement set. The success of sparse recovery relies on the knowledge of domains in which compressible representations of the image can be obtained. In this work, we consider two- and three-dimensional images, and investigate the effects of the fractional Fourier (FRT) and linear canonical transforms (LCT) in obtaining sparser transform domains. For 2D images, we investigate direct transforming versus several patching strategies. For the 3D case, we consider biomedical images, and compare several different strategies such as taking 2D slices and optimizing for each slice and direct 3D transforming.
image match has been widely used in computer vision, pattern recognition and imageprocessing. The matching efficiency is a focus topic in the field and some methods have been presented, such as simplification of simi...
详细信息
ISBN:
(纸本)9781538606971
image match has been widely used in computer vision, pattern recognition and imageprocessing. The matching efficiency is a focus topic in the field and some methods have been presented, such as simplification of similarity measure, application of optimization algorithms. Particle swarm optimization algorithm (PSO) has been utilized successfully for image match. However, it is easy to fall into the local optimum and the accuracy isn't good enough. The Water Wave Optimization (WWO) algorithm is a new evolutionary algorithm, which has been proved to be superior to many leading heuristic optimization algorithms on some benchmarking problems and engineering practical problems. In the paper, gray correlation analysis is used to simplify the calculation of similarity measure, and then WWO is employed to obtain the best matching position fast. Experimental results demonstrate that the proposed approach has higher efficiency and matching accuracy than GPSO and GABCA, meanwhile, it has good anti-noise performance.
Compressive sensing is a signal processing technique for efficiently acquiring and reconstructing the signals by finding solutions to underdetermined linear systems. Non- linear optimization algorithms are used to sol...
详细信息
ISBN:
(纸本)9781509061068
Compressive sensing is a signal processing technique for efficiently acquiring and reconstructing the signals by finding solutions to underdetermined linear systems. Non- linear optimization algorithms are used to solve underdetermined linear system. Greedy algorithms are widely used due to their low complexity than that of non-linear optimization algorithms. The problem with the Greedy algorithms is that some algorithms reduces reconstruction time but quality of reconstruction is affected while some require more reconstruction time with high quality of reconstruction. This is serious issue for the large size images. In this paper we have worked on CoSaMP algorithm and by modifying it we have reduced reconstruction time to reconstruct the sparse image with high quality of reconstruction which solves the above mention problem. Experimentation is performed on Modified CoSaMP using orthogonal filters like db4 , Haar and coif3 for different size of the images and results in terms of PSNR (peak signal to noise ratio), SSIM (Structural similarity index measurement) and runtime have been calculated and compared with OMP, ROMP, OLS and CoSaMP.
This paper gives comparative analysis of parallel morphological filter with previously designed morphological filters. In this paper, implementation is done for gray image using various structuring elements. In past, ...
详细信息
ISBN:
(纸本)9781538608142
This paper gives comparative analysis of parallel morphological filter with previously designed morphological filters. In this paper, implementation is done for gray image using various structuring elements. In past, multiple systems have been developed that have huge latency and fewer throughputs. This system reduced latency and improved through by using parallel approach. It provides a programmable and efficient implementation of basic morphological operators such as Dilation and Erosion using efficient parallel technique for different applications. It reduces processing time with respect to image dimension. Here performance analysis is based on different shape and size of structuring elements which support parallelism. MATLAB R2013a and Xilinx Design Suite 13.1 ISE are used for synthesizing this architecture that is verified by using Xilinx ISIM Simulator and prototyped on Spartan 3E FPGA Board. The proposed architecture is examined on different gray scale images.
This system was made due to frequent errors in performing calculation of goods manufactured manually. The issues can be solved by using a TCS3200 sensor that can identifies the color of goods based on the color that h...
详细信息
ISBN:
(纸本)9781538669228
This system was made due to frequent errors in performing calculation of goods manufactured manually. The issues can be solved by using a TCS3200 sensor that can identifies the color of goods based on the color that has adjusted. The use of this sensor can be a solution that will help increase the productivity of companies. In this paper, it is described the development of color sensor that is used to sort items by color in the industrial world. Technology of this control system assist in making the Internet of Things products, as well as the mechanical systems in the form of a robot arm which consists of four micro servos are designed to pick up and put stuff in the container. The result of the arithmetic processing of goods displayed through a Web page that has been designed. The achievement is using a micro controller NodeMCU in programming sensor to sort items and adjust into the container. It is also included a solution to the color sensor to recognize colors using search algorithms.
In this paper, a color digital image sharpening method is developed by using discrete sine transform (DST) and matrix low-pass Butterworth filters. First, the DST method is used to design matrix Butterworth filters. T...
详细信息
ISBN:
(纸本)9781509040179
In this paper, a color digital image sharpening method is developed by using discrete sine transform (DST) and matrix low-pass Butterworth filters. First, the DST method is used to design matrix Butterworth filters. Then, the un-sharp masking method and the designed matrix low-pass Butterworth filter are employed to develop a color image sharpening algorithm. Finally, Lena digital color image is used to demonstrate the usefulness of the proposed sharpening approach.
暂无评论