Showing papers on "3D single-object recognition published in 1986"

An Autoregressive Model Approach to Two-Dimensional Shape Classification

Abstract: In this paper, a method of classifying objects is reported that is based on the use of autoregressive (AR) model parameters which represent the shapes of boundaries detected in digitized binary images of the objects. The object identification technique is insensitive to object size and orientation. Three pattern recognition algorithms that assign object names to unlabelled sets of AR model parameters were tested and the results compared. Isolated object tests were performed on five sets of shapes, including eight industrial shapes (mostly taken from the recognition literature), and recognition accuracies of 100 percent were obtained for all pattern sets at some model order in the range 1 to 10. Test results indicate the ability of the technique developed in this work to recognize partially occluded objects. Processing-speed measurements show that the method is fast in the recognition mode. The results of a number of object recognition tests are presented. The recognition technique was realized with Fortran programs, Imaging Technology, Inc. image-processing boards, and a PDP 11/60 computer. The computer algorithms are described.

Recognizing partially visible objects using feature indexed hypotheses

Abstract: A common task in computer vision is to recognize the objects in an image Most computer vision systems do this by matching models for each possible object type in turn, recognizing objects by the best matches This is not ideal, as it does not take advantage of the similarities and differences between the possible object types The computation time also increases linearly with the number of possible objects, which can become a problem if the number is large This paper describes a new recognition method, the feature indexed hypotheses method, which takes advantage of the similarities and differences between object types, and is able to handle cases, where there are a large number of possible object types, in sub-linear computation time A two-dimensional occluded parts recognition system using this method is described

A general moment-invariants/attributed-graph method for three-dimensional object recognition from a single image

Abstract: A consistent development of general moment invariants of affine transformations for two-dimensional image functions is presented. Based on this development, a new general moment-invariants/attributed-graph (MIAG) method is presented for the identification of three-dimensional objects from a single observed image using a model-matching approach. The three-dimensional location and orientation parameters of the object are also obtained as a byproduct of the matching procedure. The scheme presented allows the observed object to be partially Occluded. For identification purposes, a three-dimensional object is represented by an attributed graph describing the geometrical structure and shape of the surface bounding the object. In such a description, two-dimensional general moment invariants of the rigid planar patches (RPP) constituting the object faces are used as attributes or feature vectors which are invariant under three-dimensional motion. With this representation, the identification problem becomes a subgraph isomorphism problem between the observed image and a library model. An algorithm is presented for this matching process, and the results are illustrated by computer simulations.

An Approach to Object Recognition: Aligning Pictorial Descriptions

Abstract: This paper examines the problem of shape-based object recognition and proposes a new approach, the alignment of pictorial descriptions. The first part of the paper reviews general approaches to visual object recognition and divides these approaches into three broad classes: invariant properties methods, object decomposition methods, and alignment methods. The second part presents the alignment method. In this approach the recognition process is divided into two stages. The first determines the transformation in space that is necessary to bring the viewed object into alignment with possible object-models. The second stage determines the model that best matches the viewed object. The proposed alignment method also uses abstract description, but unlike structural description methods, it uses them pictorially, rather than in symbolic structural descriptions.

Object recognition using tactile image array sensors

Abstract: The objective of this paper is to develop an object recognitlon system through the combination of 2-D tactile image array and visual sensors. A video camera is used to acquire a top view image of an object and two tactile sensing arrays mounted on a gripper are used to detect the tactile information about the lateral surfaces of the object. 3-D reference object models are established as a decision tree, and recognition of unknown objects is accomplished through measuring and comparing input object features hierarchically with these of the reference objects associated with the decision tree. The clustering process and recognition procedures are described. The recognition scheme has been implemented. The resulting decision tree is also presented.

Tactile recognition by probing: identifying a polygon on a plane

Abstract: An outstanding problem in model-based recognition of objects by robot systems is how the system should proceed when the acquired data are insufficient to identify uniquely the model instance and model pose that best interpret the object In this paper, we consider the situation in which some tactile data about the object are already available, but can be ambiguously interpreted The problem is thus to acquire and process new tactile data in a sequential and efficient manner, so that the object can be recognised and its location and orientation determined An object model, in this initial analysis of the problem, is a polygon located on a plane; the case of planar objects presents some interesting problems, and is also an important prelude to recognition of three-dimensional (polyhedral) objects

A Syntactic Approach to Three-Dimensional Object Recognition

Abstract: A syntactic approach to three-dimensional object recognition from a single viewpoint is proposed. The system consists of two major parts: analysis and recognition. The analysis part consists of selecting primitive surface patches and modeling grammar construction. The recognition part consists of preprocessing, image segmentation, visible primitive surfaces identification, camera model estimation, and structural analysis. Attention is given primarily to the analysis and the structural analyzer of this system.

Development of a High-Performance Stamped Character Reader

Abstract: A high-performance optical character reader (OCR) capable of reading low-quality stamped alphanumeric characters, on a metal rod, is described in this paper. Reliable and precise recognition is achieved by means of a special scanner and dedicated recognition module. The scanner revolves around the metal rod and detects characters, thus avoiding rotation of the heavy and long rod. Diffused illumination which produces a clear image, a powerful pattern-matching method?the Multiple Similarity Method (MSM)?implemented by the recognition module, and the use of adaptive rescan control results in a correct recognition rate of 99.93 percent. One recognition module can process images from up to five different scanners. The processing speed is about 1.5 s/rod. The design and industrial application of the system are discussed.

Recognition by components: a theory of visual pattern recognition

Abstract: Publisher Summary This chapter describes recent research and theory on the human's ability to recognize visual entities. The fundamental problem of object recognition is that any single object can project an infinity of image configurations to the retina. The orientation of the object to the viewer can vary continuously, each giving rise to a different two-dimensional projection. The object can be occluded by other objects or texture fields, as when viewed behind foliage. In addition, the object can even be missing some of its parts or be a novel exemplar of its particular category. However, it is only with rare exceptions that an image fails to bse rapidly and readily classified, either as an instance of a familiar object category or as an instance that cannot be so classified.

Range image segmentation for object recognition

Abstract: Segmentation is a fundamental first step in recognition of objects in range images, providing region and boundary delineation for image analysis. The simplest boundaries of interest in a range image are step edges, indicated by discontinuities in range. These points are detected by finding directed local maxima in the gradient magnitude, normalized over a small neighborhood. More complex are roof edge boundaries, indicated by a sharp bend in the surface. Sharp bends in the surface along roofs, creases, and at corners are well represented by root-mean-square (rms) curvature, a scalar measure of the surface curvature. Roof edge points are determined by finding a local peaks in the estimated rms curvature. The effect of noise in the image on roof boundary estimation accuracy is lessened by adaptively filtering the range image prior to surface fitting and curvature calculation. A new technique is presented for closing gaps in the boundary map by dilating the boundaries in a reversible manner during the segmentation process. A chamfer map is generated indicating the distance of each pixel from an edge. The image is segmented using an externally specified radius and the chamfer map to close gaps of a small, but arbitrary size. The dilated boundaries are relabeled after segmentation to maximize region area, regaining information lost to the dilated edges for later processes.

Object recognition from range images (vision, evidence)

Abstract: The recognition of objects in 3-dimensional space is an essential capability of the ideal computer vision system. Range images directly measure 3D surface coordinates of the visible portion of a scene and are well suited for this task. We report a procedure to identify 3D objects in range images, which makes use of four key processes. The first process segments the range image into "surface patches" by a clustering algorithm using surface points and associated surface normals. The second process classifies these patches as planar, convex, or concave based on a nonparametric statistical test for trend. The third process derives patch boundary information, and the results of this and the second process are used to merge compatible patches to produce reasonable object faces. The fourth process takes the patch and boundary information provided by the earlier stages and derives a representation of the range image. A list of salient features of the various objects in the database forms the core of an object recognition system, which looks for instances of these features in the representation. Occurrences of these salient features are interpreted as evidence for or against the hypothesis that a given object occurs in the scene. A measure of similarity between the set of observed features and the set of salient features for a given object in the database is used to determine the identity of an object in the scene or reject the object(s) in the scene as unknown. This evidence-based object recognition system correctly identified objects in 30 out of 31 range images. Four range images showing objects not included in the object database were rejected, as desired. Recognition degraded only slightly when evidence weights were perturbed.

From two-view motion equations to three-dimensional motion parameters and surface structure: A direct and stable algorithm

Abstract: Given a set of corresponding points from two perspective projection images of a moving rigid object, this paper presents a direct and stable algorithm to solve for the parameters of the moving object. This involves determining the rotation, the translation direction and the relative depths of object points. Unlike previous algorithms (see [1]), the current algorithm does not require determining the mode of motion, i.e. if or not the motion is a pure rotation. As easily seen, it is hard or even impossible to determine the mode of motion in the presence of noise,

An approach to 3-D object identification using range images

Abstract: The problem of speedily and reliably interpreting range data for industrial object recognition is becoming critically important in the field of robotics and computer vision. This paper presents one approach to attack this problem. It attempts to perform as much classification analysis as possible during the off-line learning process, while only a very small subset of the range data needs to be processed for the on-line recognition. An adaptive matched filter is developed to extract an object's surface feature vectors using range measurement and surface normals. A sparse representation for each object category j is constructed for classification purpose in the form of a Rj-table of selected feature vectors. The approach, based on the concept of the generalized Hough Transform, classifies an object by examining the maximum votes which it receives from various Ri-tables. An optimal selection rule is established for minimizing the misclassification probability. An experiment with a set of simulated range images of nine categories demonstrated the success of the proposed methodology.

Integrating Sensory Data for Object Recognition Tasks

Abstract: Object recognition is a difficult task for single sensor systems (e.g. machine vision) in unconstrained environments. A useful approach is to combine sensory data from more than one source to overcome these problems. However, using multiple sensors poses new problems with respect to coordination of the sensors, strategies for their use and integration of their data. In this paper, these problems are explored and solutions posed for the task of object recognition using passive stereo vision and active tactile sensing.

A theorem proving based pattern recognition system

Abstract: An object recognition system has been developed which incorporates topological as well as geometric information to match viewpoint dependent object descriptors. Theorem proving techniques are used to produce symbolic pattern matches. The recognition process uses a three phase approach. First, hypotheses are generated which correspond to model descriptors that are likely to match the data. Evidence is applied to viable hypotheses to produce a partial match. The partial match is then used to constrain the full recognition process which leads to object identification. This strategy has been found to strongly constrain the search space of possible matches and leads to large reductions in recognition times. The major contributions of the system are the representation scheme and the use of theorem proving techniques to verify object identities. This approach permits describing objects at a variety of levels and facilitates recognition despite missing information or the inclusion of artifactual data. Results of the recognition process on synthetic and actual laser range data are presented for curved and planar objects. The system is shown to operate with robustness and alacrity.

Object recognition sensor using ring pattern method for robotic applications

Abstract: To provide robots with a flexibility of operation, it is necessary to develop a sensor which can recognize the distance to the object as well as its surface shape. This paper investigates the robot sensor using the ring pattern method, where a circular pattern is projected on the object and the location and the shape of the object is recognized by analyzing the distortion of the projected pattern. In the proposed system, a conical beam is projected from a ring projector, and by varying the azimuth angle of the cone, both the distance and the shape are recognized. The distance to the object is determined utilizing the geometrical relation when the beam is focused to a point on the object surface. For the shape, the ring pattern is approximated by ellipses, and the shape is analyzed by the distributions of this parameter. The three-dimensional data are obtained from the distortion of the projected pattern from the circle, which would have been obtained by projecting the light on a plane which is parallel to the projector plane and passing through the point of the measured distance. The measurement is made by scanning the beam on the object, and the result of recognition for the whole object is represented by the surface patching. An experimental system was constructed based on the proposed principle, using 6 projectors. A satisfactory result was obtained for the measurements of distance, shape and inclination of the plane. Using a book as the object, a grid scanning is made and the result is represented by the surface patching. It is seen that the edge and the curve of the paper can be recognized.

Connected word recognition by overlap and split of reference patterns and its performance evaluation tests

Abstract: This paper proposes a new method for connected word recognition. To reduce pattern matching errors which are caused by coarticulation, talking rate variation, and silence between words in continuous speech, overlap and splitting of reference patterns and normalization of accumulated distances are introduced into our basic recognition algorithm. Also new algorithms to reduce the amount of computation and to train the recognition system are proposed. Performance evaluation tests show the efficiency of these algorithms and the entire method. Dominant factors of speech data influencing recognition performance are also investigated.

Theory and implementation of a high capacity 3-D recognition system

Abstract: A 3-D object recognition procedure employing parallel stripe projection is being developed. The systems characteristic are based on implementation in an industrial environment where high recognition rates over a long period of operation, as well as high speed, are first priorities. The recognition scheme concentrates on a structural interpretation based on the 3-D positions of planes, as well as on retrieval and encoding of surface data. An approach that seems to be most promising for this purpose is one that brings an arbitrarily positioned surface to its standard view. Model objects, as well as real objects, are studied for the purpose of evaluating the robustness as well as expertise of the algorithm employed.

Some Issues on Object Representation in Computer Vision.

Abstract: : It is argued that for most objects in reasonably unconstrained domains previous representations in computer recognition will fail due to the great diversity of appearances of objects within their class. A chair is considered in some detail to illustrate the point. To overcome problems of object variability and other hurdles in recognition a novel type of representation is introduced. This encompasses not only the familiar spatial information but include data on the function and context of what being recognised. This approach is described and some ideas are given on a 'paper' or hypothetical implementation. (Author)

Contour recognition: A syntax-directed approach

Abstract: Recognition of the outlines of objects is a fundamental problem in image analysis, with applications in many areas, including industrial inspection and medical imaging. This paper describes how an optimal method of edge analysis can be used in conjunction with a syntax describing the outlines of objects to give a method of object recognition that allows knowledge of the object to be included in the contour analysis.