The article proposes methods for increasing the efficiency of pattern recognition in images using convolutional neural networks in conditions of insufficient reference images. It is proposed to use various kinds of im...
详细信息
ISBN:
(数字)9781728160726
ISBN:
(纸本)9781728160733
The article proposes methods for increasing the efficiency of pattern recognition in images using convolutional neural networks in conditions of insufficient reference images. It is proposed to use various kinds of image transformations to generate new images and increase the training sample volume. In particular, methods of scaling, adding noise, blurring, etc. are proposed. In addition to standard image transformations, it is supposed to increase the database by presenting patterns using a doubly stochastic model. In addition, the need for the use of regularization in training is noted. The recognition results are compared for two convolutional neural networks that differ only in the number of neurons for different initial data. It is shown that the highest recognition accuracy is provided by a network trained on a large number of augmentations in combination with images generated by a doubly stochastic model.
Modified Gram-Schmidt (MGS) algorithm is one of the most-known forms of QR decomposition (QRD) algorithms. It has been used in many signal and imageprocessing applications to solve least square problem and linear equ...
详细信息
Modified Gram-Schmidt (MGS) algorithm is one of the most-known forms of QR decomposition (QRD) algorithms. It has been used in many signal and imageprocessing applications to solve least square problem and linear equations or to invert matrices. However, QRD is well-thought-out as a computationally expensive technique, and its sequential implementation fails to meet the requirements of many real-time applications. In this paper, we suggest a new parallel version of MGS algorithm that uses vLIW (very Long Instruction Word) resources in an efficient way to get more performance. The presented parallel MGS is based on compact vLIW kernels that have been designed for each algorithm step taking into account architectural and algorithmic constraints. Based on instruction scheduling and software pipelining techniques, the proposed kernels exploit efficiently data, instruction and loop levels parallelism. Additionally, cache memory properties were used efficiently to enhance parallel memory access and to avoid cache misses. The robustness, accuracy and rapidity of the introduced parallel MGS implementation on vLIW enhance significantly the performance of systems under severe rea-time and low power constraints. Experimental results show great improvements over the optimized vendor QRD implementation and the state of art.
In this article, we use among and the best-known library is Open Computer vision we call it for short OpenCv. It is used for imageprocessing, to do all operations we want, to isolate and detect a specific object, whi...
详细信息
ISBN:
(数字)9781728166544
ISBN:
(纸本)9781728166551
In this article, we use among and the best-known library is Open Computer vision we call it for short OpenCv. It is used for imageprocessing, to do all operations we want, to isolate and detect a specific object, which in our case are traffic road signs. We process to find the most efficient methods of detection of traffic road signs. Our objective is to demonstrate the links the elements for optimized and powerful to computer vision algorithms that are easy to use as typing in an image and video processing. Most of the techniques they employed the color selection, edge detection, a region of interest selection, and shapes transformation of detection. Many applications require the recognition of traffic road signs in urban areas. The automation of this task is necessary, for example, ADAS systems, and the application of vision robotics or an autonomous vehicle, consist of recognizing and identifying traffic road signs as quickly as possible with errors to be minimized, in images of their type acquired by an embedded camera on board a vehicle.
This paper presents the use of hyperspectral imageprocessing as an alternative to traditional mineral exploration techniques for identification of regions in the study area rich in zinc mineral. Zincian Dolomite is a...
详细信息
This paper is devoted to finding the source images for the processed images of social networks. Existing methods and approaches that take place in the performance of this task are considered. The algorithms of image r...
详细信息
In recent years, deep neural networks have exhibited numerous advantages in hyperspectral image classification (HIC). However, owing to the limited number of training samples of hyperspectral images (HSIs), the networ...
详细信息
In recent years, deep neural networks have exhibited numerous advantages in hyperspectral image classification (HIC). However, owing to the limited number of training samples of hyperspectral images (HSIs), the network structure should not be designed too deep to retard the overfitting phenomenon. This study proposes a cascaded dual-scale crossover network for HIC, which not only could extract rich features, but also does not make the network deeper. It continuously connects two different cascaded dual-scale crossover blocks, and automatically extracts the spectral-spatial features of HSIs. Moreover, for the limited training samples, the proposed network could flexibly capture more discriminant contextual features by using different spectral-size and spatial-size convolution kernels. Furthermore, two different cross-merge methods are designed to improve the information flow and contrast of the images to obtain parts of interest for the images. Two skip structures are also used for alleviating overfitting and accelerating the network training. Additional experimental results on some datasets, including Indian Pines, Kennedy Space Center, and University of Pavia, verify the feasibility of the proposed network. Namely, the classification accuracy of the proposed network is superior to that of other existing networks. (C) 2019 Elsevier B.v. All rights reserved.
The article deals with a problem of three-dimensional crystal lattice reconstruction, which is an important stage in the X-ray structural analysis. The accuracy of parametric and structural identification of crystals ...
详细信息
The article deals with a problem of three-dimensional crystal lattice reconstruction, which is an important stage in the X-ray structural analysis. The accuracy of parametric and structural identification of crystals directly depends on the quality of crystal lattice reconstruction. The proposed algorithm of reconstruction of a three-dimensional crystal lattice is based on minimizing the distances from each node to a line projected onto a specified plane. Three sets of two-dimensional node coordinates, obtained from three two-dimensional projections, are used as input data. We performed an analytical calculation of the reconstruction error, allowing the total reconstruction accuracy to be estimated. The results of computational experiments confirmed the high quality of the proposed reconstruction algorithms and its stability against the distortion of node coordinates. In addition, we revealed a problem of lattice system separability, with the identification accuracy for monoclinic, rhombic and tetragonal systems found to be 34%, 53% and 10%, respectively.
In this paper, we first analyze the accuracy of 3D object reconstruction using point cloud filtering applied on data from a RGB-D sensor. Point cloud filtering algorithms carry out upsampling for defective point cloud...
详细信息
Robust perception algorithms are a vital ingredient for autonomous systems such as self-driving vehicles. Checking the correctness of perception algorithms such as those based on deep convolutional neural networks (CN...
详细信息
ISBN:
(纸本)9783981926323
Robust perception algorithms are a vital ingredient for autonomous systems such as self-driving vehicles. Checking the correctness of perception algorithms such as those based on deep convolutional neural networks (CNN) is a formidable challenge problem. In this paper, we suggest the use of Timed Quality Temporal Logic (TQTL) as a formal language to express desirable spatio-temporal properties of a perception algorithm processing a video. While perception algorithms are traditionally tested by comparing their performance to ground truth labels, we show how TQTL can be a useful tool to determine quality of perception, and offers an alternative metric that can give useful information, even in the absence of ground truth labels. We demonstrate TQTL monitoring on two popular CNNs: YOLO and SqueezeDet, and give a comparative study of the results obtained for each architecture.
Space-filling curves are well known for preserving pixel locality when they are used as paths to traverse 2D image data. Some prediction-based compression algorithms make use of these curves to ensure high pixel value...
详细信息
ISBN:
(纸本)9781728140698
Space-filling curves are well known for preserving pixel locality when they are used as paths to traverse 2D image data. Some prediction-based compression algorithms make use of these curves to ensure high pixel values correlation during 2D image data traversal. This work explores the distribution of pixel correlation induced by all possible space-filling curves on 2D image data and demonstrates that commonly used curves, such as the Hilbert or the Peano curves, do not provide the best possible pixel correlation for natural photographic images. Using experimental data collected on a large set of such images, we demonstrate that row-prime ordering is the best choice for preserving maximum pixel values correlation while reducing the dimensionality of 2D natural photographic image data.
暂无评论