Search results for "processing"
showing 10 items of 8572 documents
A word prediction methodology for automatic sentence completion
2015
Word prediction generally relies on n-grams occurrence statistics, which may have huge data storage requirements and does not take into account the general meaning of the text. We propose an alternative methodology, based on Latent Semantic Analysis, to address these issues. An asymmetric Word-Word frequency matrix is employed to achieve higher scalability with large training datasets than the classic Word-Document approach. We propose a function for scoring candidate terms for the missing word in a sentence. We show how this function approximates the probability of occurrence of a given candidate word. Experimental results show that the proposed approach outperforms non neural network lang…
Spatio-Temporal Saliency Detection in Dynamic Scenes using Local Binary Patterns
2014
International audience; Visual saliency detection is an important step in many computer vision applications, since it reduces further processing steps to regions of interest. Saliency detection in still images is a well-studied topic. However, videos scenes contain more information than static images, and this additional temporal information is an important aspect of human perception. Therefore, it is necessary to include motion information in order to obtain spatio-temporal saliency map for a dynamic scene. In this paper, we introduce a new spatio-temporal saliency detection method for dynamic scenes based on dynamic textures computed with local binary patterns. In particular, we extract l…
Content based segmentation of patterned wafers
2004
We extend our previous work on the image segmentation of electronic structures on patterned wafers to improve the defect detection process on optical inspection tools. Die-to-die wafer in- spection is based on the comparison of the same area on two neigh- boring dies. The dissimilarities between the images are a result of defects in this area of one of the dies. The noise level can vary from one structure to the other, within the same image. Therefore, seg- mentation is required to create a mask and apply an optimal thresh- old in each region. Contrast variation on the texture can affect the response of the parameters used for the segmentation. We show a method to anticipate these variation…
Pattern image enhancement by extended depth of field
2014
Abstract Most optical defect localization techniques such as dynamic laser stimulation or photon emission microscopy require a pattern image of the device to be taken. The main purpose is for device navigation, but it also enables the analyst to identify the location of the monitored activity by superimposing it onto the pattern image. The defect localization workflow usually starts at low or medium magnification. At these scales, several factors can lead to a lack of orthogonality of the sample with the optical axis of the system. Therefore, images can be locally out of focus and poorly resolved. In this paper, a method based on Depth of Field Extension is suggested to correct the pattern …
Modified LACIF filtering in background disjoint noise
2011
Abstract This work deals with pattern recognition methods based on correlations for images in the presence of noise. We propose a modification of the nonlinear Locally Adaptive Contrast Invariant Filter (LACIF) that yields correlation peaks that are invariant to linear intensity changes of the target but that has some limitations in the presence low variance nonoverlapping background noise. The modification of the filter implies a normalization by a global variance of several distributions. The estimation of the variance distributions is done locally by means of correlations. Experimental results as well as comparisons with the classical matched filter and the common LACIF are given.
Intensity invariant nonlinear correlation filtering in spatially disjoint noise.
2010
We analyze the performance of a nonlinear correlation called the Locally Adaptive Contrast Invariant Filter in the presence of spatially disjoint noise under the peak-to-sidelobe ratio (PSR) metric. We show that the PSR using the nonlinear correlation improves as the disjoint noise intensity increases, whereas, for common linear filtering, it goes to zero. Experimental results as well as comparisons with a classical matched filter are given.
Perceptually weighted optical flow for motion-based segmentation in MPEG-4 paradigm
2000
In the MPEG-4 paradigm, the sequence must be described in terms of meaningful objects. This meaningful, high-level representation should emerge from low-level primitives such as optical flow and prediction error which are the basic elements of previous-generation video coders. The accuracy of the high-level models strongly depends on the robustness of the primitives used. It is shown how perceptual weighting in optical flow computation gives rise to better motion estimates which consistently improve motion-based segmentation compared to equivalent unweighted motion estimates.
View Planning Approach for Automatic 3D Digitization of Unknown Objects
2012
International audience; This paper addresses the view planning problem for the digitization of 3D objects without prior knowledge on their shape and presents a novel surface approach for the Next Best View (NBV) computation. The proposed method uses the concept of Mass Vector Chains (MVC) to define the global orientation of the scanned part. All of the viewpoints satisfying an orientation constraint are clustered using the Mean Shift technique to construct a first set of candidates for the NBV. Then, a weight is assigned to each mode according to the elementary orientations of its different descriptors. The NBV is chosen among the modes with the highest weights and which comply with the rob…
Combining similarity measures in content-based image retrieval
2008
The purpose of content based image retrieval (CBIR) systems is to allow users to retrieve pictures from large image repositories. In a CBIR system, an image is usually represented as a set of low level descriptors from which a series of underlying similarity or distance functions are used to conveniently drive the different types of queries. Recent work deals with combination of distances or scores from different and usually independent representations in an attempt to induce high level semantics from the low level descriptors of the images. Choosing the best method to combine these results requires a careful analysis and, in most cases, the use of ad-hoc strategies. Combination based on or…
Optical flow estimation from multichannel spherical image decomposition
2011
The problem of optical flow estimation is largely discussed in computer vision domain for perspective images. It was also proven that, in terms of optical flow analysis from these images, we have difficulty distinguishing between some motion fields obtained with little camera motion. The omnidirectional cameras provided images with large filed of view. These images contain global information about motion and allow to remove the ambiguity present in perspective case. Nevertheless, these images contain significant radial distortions that is necessary to take into account when treating these images to estimate the motion. In this paper, we shall describe new way to compute efficient optical fl…