The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
In this paper we propose a deep learning architecture to make the best use of global and local information for pixel-wise semantic segmentation. The architecture of three-skips CNN is built with convolutional layers in VGG16 network and its mirrored convolutional layers. Our architecture aims to road scene understanding. In order to save memory and computational time, we use unpooling layers to map...
The hyperspectral remote sensing realizes image formation of spatial feature of the target by taking nanoscale super-high spectral resolution and meanwhile performs continuous spectral covering for dozens of and even hundreds of narrow wave bands formed on the wider wavelength range of each image element, which is with the ability of distinguishing the subtle difference of ground object spectrum,...
In view of old-fashioned digital display instrument of substation, an improved character recognition method for digital display instrument is proposed. It is a method mainly based on pulse coupled neural network which comes from study of cat's visual cortex neurons. Apart from synchronous oscillation that is one characteristic of pulse coupled neural network being used in image segmentation, the other...
Traditional coded structured light with spatial neighborhood is a well-known technique which is used to acquire 3D information. In this paper, we use stereo structure by adding a camera so that the calibration of projector could be omitted. By projecting a new color stripe pattern in which the thick stripes are encoded of De Bruijn sequences with six different channel values and order three, while...
This paper introduces an image fusion technique on a gated image. Gated imaging is a powerful technique to capture images in turbid condition. By controlling the time delay of the shutter to capture the returning light, we can obtain the image at different depth. Modular transfer function (MTF) is used to evaluate the quality of the gated images. Images with useful information (clear and sharp objects)...
This paper presents a method of insect recognition using computer vision technology. First, we extracted fourteen features from images of some species of insects. These features are rectangularity, elongation, roundness, eccentricity, sphericity, lobation, compactness and seven Hu moment invariants. Second, a machine learning algorithm named Random Trees was employed, to play a role of a classifier...
The goal of the study is to develop automatic fabric analysis system by using inexpensive image processing techniques. In this study, we proposed a novel automatic method for woven fabric structure identification. This method is based on widely used digital image analysis techniques. It allows automatic weft yarn and warp yarn cross area segmentation through a spatial domain integral projection approach...
In this paper, we present a novel automatic method for woven textile structure recognition in fiber-level. This method is based on digital image analysis techniques. It allows automatic weft yarn and warp yarn crossed-area segmentation through a spatial domain integral projection approach. Secondly, by applying unsupervised fuzzy c-means clustering on multi-scale direction invariant texture features...
In view of the speckle noise in the SAR images, utilizing the Contourlet's advantages of multiscale, localization, directionality and anisotropy, a new SAR image fusion segmentation algorithm based on the persistence and clustering in the Contourlet domain is proposed in this paper. The algorithm captures the persistence and clustering of the Contourlet transform, which is modeled by HMT and MRF,...
Traditional fuzzy Hopfield neural network (FHNN) is one of the excellent segmentation methods for CT image. Although FHNN has the capacity of searching values with high precision, it has obvious disadvantages, such as local minimum and slow convergence. In order to make up these shortcomings and find the right global minimum, a FHNN Algorithm based on genetic approach is proposed. Fine segmentation...
An improved efficient fractal algorithm, based on higher-order statistics (HOS), is presented for infrared (IR) small target detection under complex background of a single image. This algorithm is divided into two parts: coarse location and fine location. It firstly uses higher-order statistics to locate the target coarsely, and then a region of interest (ROI) containing the infrared small target...
Video-on-demand (VoD) service has been greatly enhanced with the maturation of both peer-to-peer (P2P) and network coding (NC) technology. Digital video cassette recording (VCR) function provides users with more freedom that they can choose different playback modes while viewing video clips. Besides, users are able to switch between different scenes with the support of video-segmentation. Based on...
In order to make the data responding to the edges of moving objects expand to the area, whose gray level is flat, it needs to keep iterative numbers large enough on gradient-based optical flow algorithm. The traditional approaches need a large amount of calculation, but results accuracy is low. The gradient number accuracy and reliability will directly determine the result of optical flow algorithm...
VCR functionality is difficult to be implemented in a P2P VoD system, since it may give more server workload and fails some distribution paths. In this paper, we first analyze the user behavior of VCR interaction, and propose a video segmentation-aided P2P VoD System with VCR functionality. The simulation results show that our mechanism could reduce the server workload and improve the VCR performance...
A novel algorithm for interactive multilabel image/video segmentation is proposed in this paper. Given a small number of pixels with user-defined (or pre-defined) labels, our method can automatically propagate those labels to the remaining unlabeled pixels through an iterative procedure. Theoretical analysis of the convergence property of this algorithm is developed along with the corresponding connections...
Image mosaic is a very important step in big frame image measurement, its precision has a great effect on the measurement precision. Twice matching can make the mosaic precision beyond the pixel dimensions. And the pixels' gray rebuilding based on pixel position and pixel area distribution also has a great help to raise the measurement precision
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.