This paper presents a comparative study and survey of model-based object-recognition algorithms for robot vision. The goal of these algorithms is to recognize the identity, position, and orientation of randomly oriented industrial parts. In one form this is commonly referred to as the "bin-picking" problem, in which the parts to be recognized are presented in a jumbled bin. The paper is organized according to 2-D, 2½-D, and 3-D object representations, which are used as the basis for the recognition algorithms. Three central issues common to each category, namely, feature extraction, modeling, and matching, are examined in detail. An evaluation and comparison of existing industrial part-recognition systems and algorithms is given, providing insights for progress toward future robot vision systems.

Model-based recognition in robot vision

The human face is an object that is easily located in complex scenes by infants and adults alike. Yet the development of an automated system to perform this task is extremely challenging. An attempt to solve this problem raises two important issues in object location. First, natural objects such as human faces tend to have boundaries which are not exactly described by analytical functions. Second, the object of interest (face) could occur in a scene in various sizes, thus requiring the use of scale independent techniques which can detect instances of the object at all scales. Although, the task of identifying a well-framed face (as one of a set of labeled faces) has been well researched, the task of locating a face in a natural scene is relatively unexplored. We present a computational theory for locating human faces in scenes with certain constraints. The theory will be validated by experiments confined to instances where people's faces are the primary subject of the scene, occlusion is minimal, and the faces contrast well against the background.

Locating human faces in photographs

An abstraction-based paradigm that makes explicit the process of imposing assumptions on data is discussed. The units of abstraction are models in which levels of abstraction are determined by the degree of assumption necessary for their application. A general-to-specific refinement process provides a mechanism to proceed gracefully through the abstraction hierarchy. This strategy was applied to the recognition and pose determination of objects comprising simple and compound cylindrical and planar surfaces in dense range data. A method of computing reliable Gaussian and mean curvature sign-map descriptors from the polynomial approximations of surfaces is demonstrated. A means for determining the pose of constructed geometric forms whose algebraic surface descriptions are nonlinear in terms of their orienting parameters is developed. It is shown that biquadratic surfaces are suitable companion-linear forms for cylinder approximation and parameter estimation. The estimates provide the initial parametric approximations necessary for a nonlinear regression stage to fine tune the estimates by fitting the actual nonlinear form to the data. >

An abstraction-based approach to 3-D pose determination from range images

This thesis advances an abstraction-based paradigm which makes explicit the process of imposing assumptions on data. The units of abstraction are models whose levels of abstraction are determined by the degree of assumption necessary for their application. A general to specific refinement process provides a mechanism to proceed gracefully through the abstraction hierarchy. The task of object recognition and pose determination becomes one of making stronger and stronger assumptions about the data. These assumptions yield model hypotheses which may then be applied and tested. This process of assumption and abstraction furnishes a path between symbolic descriptors of objects in the scene to their numeric specification. Throughout the process, abstractions are made to better interpret the original data to which the hypothesized models are fitted/computed. The strategy is thus data-bound. This strategy was applied to the recognition and pose determination of objects comprising cylindrical and planar surfaces in dense range data.
A method of computing reliable Gaussian and mean curvature sign-map descriptors from the polynomial approximations of surfaces was demonstrated. Such descriptors which are invariant under perspective variation are suitable for hypothesis generation.
A means for determining the pose of constructed geometric forms whose algebraic surface descriptions are non-linear in terms of their orienting parameters was developed. This was done by means of linear functions which are capable of approximating non-linear forms and determining their parameters. It was shown that biquadratic surfaces are suitable companion-linear forms for cylinder approximation and parameter estimation. The estimates provided the initial parametric approximations necessary for a non-linear regression stage to fine tune the estimates by fitting the actual non-linear form to the data.
A hypothesis-based split-merge algorithm for extraction and pose determination of cylinders and planes which merge smoothly into other surfaces was developed. It was shown that all split-merge algorithms are hypothesis-based.
A finite-state algorithm for the extraction of the boundaries of run-length regions was developed. The computation takes advantage of the run list topology and boundary direction constraints implicit in the run-length encoding.

/pdf/on-three-dimensional-object-recognition-and-pose-1weyjtrrhp.pdf

Thomas S Huang

Papers

Recognition of curvilinear objects by matching relational structures.