A transitory image sequence is one in which no scene element is visible through the entire sequence. This article deals with some major theoretical and algorithmic issues associated with the task of estimating structu...
详细信息
A transitory image sequence is one in which no scene element is visible through the entire sequence. This article deals with some major theoretical and algorithmic issues associated with the task of estimating structure and motion from transitory image sequences. Two representations, world-centered (WC) and camera-centered (CC), behave very differently with a transitory sequence. The asymptotical error properties derived in this article indicate that one representation is significantly superior to the other, depending on whether one uses camera-centered or world-centered estimates. Rigorous experiments were conducted with real-image sequences taken by a fully calibrated camera system. The comparison demonstrated that a good accuracy can be obtained from transitory image sequences.< >
This paper develops a new class of physics-based deformable models which can deform both globally and locally. Their global parameters are functions allowing the definition of new parameterized primitives and paramete...
详细信息
This paper develops a new class of physics-based deformable models which can deform both globally and locally. Their global parameters are functions allowing the definition of new parameterized primitives and parameterized global deformations. These new global parameter functions improve the accuracy of shape description through the use of a few intuitive parameters such as functional bending and twisting. Using a physics-based approach we convert these geometric models into deformable models that deform due to forces exerted from the data-points so as to conform to the given dataset. We present an experiment involving the extraction of shape and motion of the Left Ventricle (LV) of a heart from MRI-SPAMM data based on a few global parameter functions.< >
This paper investigates the use of morphological size distributions to characterize textures and explore their sensitivity to the presence of inhomogeneities and multiple textures. When the image texture is a result o...
详细信息
This paper investigates the use of morphological size distributions to characterize textures and explore their sensitivity to the presence of inhomogeneities and multiple textures. When the image texture is a result of material properties such variations in texture typically correspond to defects on the surface being imaged. An application where the analysis of texture variations is of great significance is the assessment of distresses on pavement surfaces. We develop a novel normalization scheme based on a Gaussian model to enhance the sensitivity of the morphological distributions to texture variations indicative of defects. Results from a simple rule-based classification scheme are presented to demonstrate that measures derived from the normalized distributions are useful in classifying distresses. Our use of normalized distributions, and not the original images, to develop measures for analyzing textures results in a significant reduction in computational and storage requirements.< >
This paper describes an automated process for the dynamic creation of a pattern-recognizing computer program consisting of initially unknown detectors, an initially-unknown iterative calculation incorporating the as-y...
详细信息
This paper describes an automated process for the dynamic creation of a pattern-recognizing computer program consisting of initially unknown detectors, an initially-unknown iterative calculation incorporating the as-yet-uncreated detectors, and an initially-unspecified final calculation incorporating the results of the as-yet-uncreated iteration. The program's goal is to recognize a given protein segment as being a transmembrane domain or non-transmembrane area. The recognizing program to solve this problem will be evolved using the recently developed genetic programming paradigm. Genetic programming starts with a primordial ooze of randomly generated computer programs composed of available programmatic ingredients and then genetically breeds the population using the Darwinian principle of survival of the fittest and the genetic crossover (sexual recombination) operation. Automatic function definition enables genetic programming to dynamically create subroutines (detectors). When cross-validated, the best genetically-evolved recognizer achieves an out-of-sample correlation of 0.968 and an out-of-sample error rate of 1.6%. This error rate is better than that recently reported for five other methods.< >
This paper presents a system for large vocabulary recognition of on-line handwritten cursive words. The system first uses a filtering module, based on simple letter features, to quickly reduce a large reference dictio...
详细信息
This paper presents a system for large vocabulary recognition of on-line handwritten cursive words. The system first uses a filtering module, based on simple letter features, to quickly reduce a large reference dictionary to a smaller number of candidates; the reduced lexicon along with the original input is subsequently fed to a recognition module. In order to exploit the sequential nature of the temporal data, we employ a TDNN-style network architecture which has been successfully used in the speech recognition domain. Explicit segmentation of the input words into characters is avoided by using a sliding window concept where the input word representation (a set of frames) is presented to the neural network-based recognizer sequentially. The outputs of the recognition module are collected and converted into a string of characters that can be matched with the candidate words. A description of the complete system and its components is given.< >
Range data offer a direct way to produce shape descriptions of surfaces. Typically single range images have the form of a graph surface z=g(z,y) and thus suffer from occlusion. One can reduce this problem by taking se...
详细信息
Range data offer a direct way to produce shape descriptions of surfaces. Typically single range images have the form of a graph surface z=g(z,y) and thus suffer from occlusion. One can reduce this problem by taking several images from different locations and merging them together. The result is a real 3-D description of the object's surface. In this paper we address several problems that result from the 2 1/2 -D to 3-D transition. We present an algorithm which is able to merge depth images of an arbitrary shaped object using a highly local approach. An explicit sensor error model is used to support the merging. A key feature of our algorithm is the ability to update its result by an additional new view. Thus, it is possible to gradually improve the surface description in regions of high noise.< >
Active and selective perception seeks regions of interest in an image in order to reduce the computational complexity associated with time-consuming processes such as object recognition. We describe in this paper a vi...
详细信息
Active and selective perception seeks regions of interest in an image in order to reduce the computational complexity associated with time-consuming processes such as object recognition. We describe in this paper a visual attention system that extracts regions of interest by integrating multiple image cues. Bottom-up cues are detected by decomposing the image into a number: of feature and conspicuity maps, while a-priori knowledge (i.e. models) about objects is used to generate top-down attention cues. Bottom-up and top-down information is combined through a non-linear relaxation process using energy minimization-like procedures. The functionality of the attention system is expanded by the introduction of an alerting (motion-based) system able to explore and avoid obstacles. Experimental results are reported, using cluttered and noisy scenes.< >
We present a framework for image registration algorithms that finds a lowest-order model of the flow between two images. Low-order models are useful in image registration, because they leave scene structure intact. Bu...
详细信息
We present a framework for image registration algorithms that finds a lowest-order model of the flow between two images. Low-order models are useful in image registration, because they leave scene structure intact. But in real images complexity varies, and cannot be determined ahead of time. Algorithms in our framework adapt model complexity to image data during a coarse-fine parameter estimation process. Complexity increases keep residual flow small enough that motion can be correctly estimated at each subsequent resolution level. We present one algorithm within this framework which increases complexity by replacing global estimates with estimates over successively smaller patches. We show results of applying this algorithm to the task of mosaicing panoramic aerial images with unknown lens distortion and unknown camera position.< >
An object recognition approach based on concurrent coarse-and-fine matching using a multi-layer Hopfield neural network is presented. The proposed network consists of several cascaded single layer Hopfield networks, e...
详细信息
An object recognition approach based on concurrent coarse-and-fine matching using a multi-layer Hopfield neural network is presented. The proposed network consists of several cascaded single layer Hopfield networks, each encoding object features at a distinct resolution, with bidirectional interconnections linking adjacent layers. The interconnection weights between nodes associating adjacent layers are structured to favor node pairs for which model translation and rotation, when viewed at the two corresponding resolutions, are consistent. This inter-layer feedback feature of the algorithm reinforces the usual intra-layer matching process in conventional single layer Hopfield nets in order to compute the model-object match which is most consistent across several resolution levels. The performance of the algorithm is demonstrated in cases of images containing single and multiple occluded objects. These results are compared with recognition results obtained using a single layer Hopfield network.< >
We present an approach to function-based object recognition that reasons about the functionality of an object's initiative parts. We extend the popular "recognition by parts" shape recognition framework ...
详细信息
We present an approach to function-based object recognition that reasons about the functionality of an object's initiative parts. We extend the popular "recognition by parts" shape recognition framework to support "recognition, by functional parts", by combining a set of functional primitives and their relations with a set of abstract volumetric shape primitives and their relations. Previous approaches have relied on more global object features, often ignoring the problem of object segmentation, and thereby restricting themselves to range images of unoccluded scenes. We show how these shape primitives and relations can be easily recovered from superquadric ellipsoids which, in turn, can be recovered from either range or intensity images of occluded scenes. Furthermore, the proposed framework supports both unexpected (bottom-up) object recognition and expected (top-down) object recognition. We demonstrate the approach on, a simple domain by recognizing a restricted class of hand-tools from 2-D images.< >
暂无评论