Most image-retrieval systems rely on similarity measures for collecting images of similar types. Similarity measures are an integral part in the development of image management systems. In this paper, we propose frame...
Most image-retrieval systems rely on similarity measures for collecting images of similar types. Similarity measures are an integral part in the development of image management systems. In this paper, we propose frame-based similarity measures for accessing structured images, e.g. images can be understood by inferring from objects present and the relationships among them. The image content is described in the following ways: (1) adjacency blocks and/or (2) unary and binary attributes that are used to fill frame slots for representing image structures. The retrieval is based on similarity measures by comparing the contents of the query image and database image. The frame-based representation scheme is application-independent. Our similarity measures allow for images to be retrieved with different degrees of similarity and are flexible. We have developed a prototype system using the paradigms proposed. We demonstrate the usefulness of our system with some experimental results. (C) 1997 Academic Press Limited.
The basic idea that the perception of actual embodied beings, be they animal or robotic, is fundamentally related to their embodiment is generally referred to as purposive or animate vision. Research in this field gen...
详细信息
The basic idea that the perception of actual embodied beings, be they animal or robotic, is fundamentally related to their embodiment is generally referred to as purposive or animate vision. Research in this field generally emphasises low-level vision techniques. This paper outlines a philosophical basis for embodied perception, and develops a framework for conceptual embodiment of vision-guided robots. The aim is to facilitate the use of high-level vision through an active perception framework. We argue that the classical computervision paradigm has problems in high-level vision due to an implicit assumption that objects in the world can be objectively subdivided into categories. Further, that through conceptual embodiment, active perception offers a way forward. We present a mobile robot navigation system based on the principles of conceptual embodiment. The system uses object recognition to guide a robot around known objects. The robot's object model is embodied, and this embodiment yields specific advantages for the robot.
We study the problem of recovering the approximate three-dimensional shape of an object when knowledge about the object is available. The application of knowledge-based methods to image processing tasks will help over...
详细信息
We study the problem of recovering the approximate three-dimensional shape of an object when knowledge about the object is available. The application of knowledge-based methods to image processing tasks will help overcome problems which arise from processing images using a pixel-based approach. We show that by applying domain specific knowledge, we can fully automate the derivation of the approximate shape of an object. Further, this approach can yield specific advantages over existing approaches, both in terms of computation and processing results. This is a powerful paradigm that will have applications in object recognition, robotic navigation, and domain specific scene understanding.
We propose the use of a Markov random field model for handwritten word recognition. The main advantage of Markov random field models is that they provide flexible and natural models for the interaction between spatial...
详细信息
ISBN:
(纸本)0780342534
We propose the use of a Markov random field model for handwritten word recognition. The main advantage of Markov random field models is that they provide flexible and natural models for the interaction between spatially related random variables in their neighborhood systems via clique functions. In our scheme, Gabor filters are adopted for feature extraction. A fuzzy neighborhood system is proposed and fuzzy matching measurements are developed to cope with the variability of handwritten word shapes. A relaxation labeling algorithm is used to maximize the global compatibilities of Markov random fields. The influence of neighborhood sizes and the iteration number on recognition rates of the system is investigated. Our initial experiments have shown encouraging results.
It is well known that a linearly separable set of classes is ideal for a pattern recognition task. The majority of pattern recognition research has been devoted to achieve linear separability of classes by nonlinear i...
详细信息
It is well known that a linearly separable set of classes is ideal for a pattern recognition task. The majority of pattern recognition research has been devoted to achieve linear separability of classes by nonlinear input-output mapping. We develop a novel idea of class label separation by projecting each element of the feature vector onto a manifold. The functional characteristics of the manifold associated with each feature type are learnt iteratively from the class label distribution under an optimization criterion. This process attempts to transform an n-dimensional nonlinearly separable feature classification task to an n-dimensional linearly separable problem. The burden of classifying features that are associated with multiple class labels is handled by projections of other discriminating features. This enables fast learning of the classification task by the second stage network which accepts the projected output as its input. If the classification task is modified by an addition of a feature element, the system requires iterative learning of the manifold associated with this new unit only and does not require learning of the whole set of features as seen in conventional neural networks. This iterative knowledge aggregation permits ease of fine tuning and selection of an optimal set of parameters for a given task. The above concept is demonstrated on a set of classification tasks.
Most data sets that describe and evolve from real-world systems are by nature semiquantitative or qualitative rather than quantitative. This can mean large variations in the significance of results that are derived fr...
详细信息
Most data sets that describe and evolve from real-world systems are by nature semiquantitative or qualitative rather than quantitative. This can mean large variations in the significance of results that are derived from this data for decision-making processes given that the original database provides training and prototypical examples that reflect systems of events in the real world. In this article we propose a structure for a Knowledge-Based System (KBS) that is derived using significance within given contextual domains. Data that would ordinarily be classified by simple attribute classification techniques are now categorized by understanding patterns and value distributions for attributes and attribute domains that exist within rich and dense databases such as in the case of census databases double dagger and Geographic Information Systems (GIS);rich by the very number of fields and interpretations, depending on the context in which the data are to be reviewed. The structure we have implemented for capturing and structuring semiquantitative information is the Fuzzy Cognitive Map (FCM). We also reduce the number of false patterns labeled ''significant'' by incorporating the knowledge used by human experts to find significance within the data. We treat this knowledge as initial background knowledge and as a minimal set for distinguishing significance for particular attribute values within a given context. (C) 1996 John Wiley & Sons, Inc.
The authors describe their system for writer independent, off-line unconstrained handwritten word recognition. First, they present a new method to automatically determine the parameters of Gabor filters to extract fea...
详细信息
ISBN:
(纸本)0780336674
The authors describe their system for writer independent, off-line unconstrained handwritten word recognition. First, they present a new method to automatically determine the parameters of Gabor filters to extract features from slant and tilt corrected images. An algorithm is also developed to translate 2D images to 1D domain. Finally, they propose a modified dynamic programming method with fuzzy inference to recognize words. Their initial experiments have shown encouraging results.
In this paper we discuss the use of covariance methods in invariant feature extraction,texture segmentation,edge detection,and surface geometry *** covariance technique is used to compute local descriptors and to inde...
详细信息
In this paper we discuss the use of covariance methods in invariant feature extraction,texture segmentation,edge detection,and surface geometry *** covariance technique is used to compute local descriptors and to index roughness,anisotropy, or general textural *** also present a simple yet effective edge detection algorithm using a neural network which is trained by invariant features generated from covariance matrices.
In this paper, we present a new off-line word recognition system that is able to recognise unconstrained handwritten words from their grey-scale images, and is based on structural information in the handwritten word. ...
详细信息
暂无评论