Image segmentation continues to be a fundamental problem in computer vision and image understanding. In this thesis, we present a Bayesian network that we use for object boundary detection in which the MPE (most probable explanation) before any evidence can produce multiple non-overlapping, non-self-intersecting closed contours and the MPE with evidence where one or more connected boundary points are provided produces a single non-self-intersecting, closed contour that accurately defines an object's boundary. We also present a near-linear-time algorithm that determines the MPE by computing the minimum-path spanning tree of a weighted, planar graph and finding the excluded edge (i.e., an edge not in the spanning tree) that forms the most probable loop. This efficient algorithm allows for real-time feedback in an interactive environment in which every mouse movement produces a recomputation of the MPE based on the new evidence (i.e., the new cursor position) and displays the corresponding closed loop. We call this interface "object highlighting" since the boundary of various objects and sub-objects appear and disappear as the mouse cursor moves around within an image. / Graduation date: 2004

We show that a color algorithm capable of separating illumination from reflectance in a Mondrian world can be learned from a set of examples. The learned algorithm is equivalent to filtering the image data---in which reflectance and illumination are mixed---through a center-surround receptive field in individual chromatic channels. The operation resembles the "retinex" algorithm recently proposed by Edwin Land. This result is a specific instance of our earlier results that a standard regularization algorithm can be learned from examples. It illustrates that the natural constraints needed to solve a problemsin inverse optics can be extracted directly from a sufficient set of input data and the corresponding solutions. The learning procedure has been implemented as a parallel algorithm on the Connection Machine System.

The computational approach to the study of vision inquires directly into the sort of information processing needed to extract important information from the changing visual image---information such as the three-dimensional structure and movement of objects in the scene, or the color and texture of object surfaces. An important contribution that computational studies have made is to show how difficult vision is to perform, and how complex are the processes needed to perform visual tasks successfully. This article reviews some computational studies of vision, focusing on edge detection, binocular stereo, motion analysis, intermediate vision, and object recognition.

This paper describes a machine vision system that classifies reflectance properties of surfaces such as metal, plastic, or paper, under unknown real-world illumination. We demonstrate performance of our algorithm for surfaces of arbitrary geometry. Reflectance estimation under arbitrary omnidirectional illumination proves highly underconstrained. Our reflectance estimation algorithm succeeds by learning relationships between surface reflectance and certain statistics computed from an observed image, which depend on statistical regularities in the spatial structure of real-world illumination. Although the algorithm assumes known geometry, its statistical nature makes it robust to inaccurate geometry estimates.

This research project is a study of the role of fixation and visual attention in object recognition. In this project, we build an active vision system which can recognize a target object in a cluttered scene efficiently and reliably. Our system integrates visual cues like color and stereo to perform figure/ground separation, yielding candidate regions on which to focus attention. Within each image region, we use stereo to extract features that lie within a narrow disparity range about the fixation position. These selected features are then used as input to an alignment-style recognition system. We show that visual attention and fixation significantly reduce the complexity and the false identifications in model-based recognition using Alignment methods. We also demonstrate that stereo can be used effectively as a figure/ground separator without the need for accurate camera calibration.

Methods for fusing two computer vision methods are discussed and several example algorithms are presented to illustrate the variational method of fusing algorithms. The example algorithms seek to determine planet topography given two images taken from two different locations with two different lighting conditions. The algorithms each employ assingle cost function that combines the computer vision methods of shape-from-shading and stereo in different ways. The algorithms are closely coupled and take into account all the constraints of the photo-topography problem. The algorithms are run on four synthetic test image sets of varying difficulty.

How the visual system extracts shape information from a single grey-level image can be approached by examining how the information about shape is contained in the image. This technical report considers the characteristic equations derived by Horn as a dynamical system. Certain image critical points generate dynamical system critical points. The stable and unstable manifolds of these critical points correspond to convex and concave solution surfaces, giving more general existence and uniqueness results. A new kind of highly parallel, robust shape from shading algorithm is suggested on neighborhoods of these critical points. The information at bounding contours in the image is also analyzed.

We describe a technique for finding pixelwise correspondences between two images by using models of objects of the same class to guide the search. The object models are 'learned' from example images (also called prototypes) of an object class. The models consist of a linear combination ofsprototypes. The flow fields giving pixelwise correspondences between a base prototype and each of the other prototypes must be given. A novel image of an object of the same class is matched to a model by minimizing an error between the novel image and the current guess for the closest modelsimage. Currently, the algorithm applies to line drawings of objects. An extension to real grey level images is discussed.

Image analysis and graphics synthesis can be achieved with learning techniques using directly image examples without physically-based, 3D models. In our technique: -- the mapping from novel images to a vector of "pose" and "expression" parameters can be learned from a small set of example images using a function approximation technique that we call an analysis network; -- the inverse mapping from input "pose" and "expression" parameters to output images can be synthesized from a small set of example images and used to produce new images using a similar synthesis network. The techniques described here have several applications in computer graphics, special effects, interactive multimedia and very low bandwidth teleconferencing.

We present a component-based approach for recognizing objects under large pose changes. From a set of training images of a given object we extract a large number of components which are clustered based on the similarity of their image features and their locations within the object image. The cluster centers build an initial set of component templates from which we select a subset for the final recognizer. In experiments we evaluate different sizes and types of components and three standard techniques for component selection. The component classifiers are finally compared to global classifiers on a database of four objects.