Subspace methods for robot vision

TLDR

Visual recognition is formulated as matching appearance rather than shape, and the set of all appearance variations defines a robot vision task’s visual workspace. The study proposes an efficient algorithm to locate the nearest point on the appearance manifold for recognition. Images are coarsely sampled from the visual workspace, compressed into a low‑dimensional eigenspace that represents the appearance manifold, and an input image is projected onto this manifold to recognize task parameters by locating its exact position. The appearance representation enables applications such as precise visual positioning, real‑time visual tracking, and real‑time temporal inspection.

Abstract

In contrast to the traditional approach, visual recognition is formulated as one of matching appearance rather than shape. For any given robot vision task, all possible appearance variations define its visual workspace. A set of images is obtained by coarsely sampling the workspace. The image set is compressed to obtain a low-dimensional subspace, called the eigenspace, in which the visual workspace is represented as a continuous appearance manifold. Given an unknown input image, the recognition system first projects the image to eigenspace. The parameters of the vision task are recognized based on the exact location of the projection on the appearance manifold. An efficient algorithm for finding the closest manifold point is described. The proposed appearance representation has several applications in robot vision. As examples, a precise visual positioning system, a real-time visual tracking system, and a real-time temporal inspection system are described.