1988•Unpublished venueRequires access

Analysis of 3D Texture

Andrew Blake, C. Marinos

Open publisher page 0 citations

Abstract

ground Several researchers have investigated the role of texture as a basis for the recovery of surface orientation. Gib-son 1 was the first to address the problem of recovering the orientation of a plane covered with textural elements. He assumed that the density (number of elements per unit area) of these elements is uniform. Under this as-sumption he observed that the density gradient in the image specifies surface orientation. More recently 2, Witkin addresses the problem of recovering the slant a and tilt r of a planar surface viewed under orthographic projection. Let /? be the angle which a tangent at a point of a surface marking makes with the projection of a fixed coordinate axis of the image plane and a the angle the projection of that tangent makes with the coor-dinate axis. Under orthographic projection then Witkin develops a geometric model which relates a to /?: a — T + arctan /tan/A \\ COS <7 / Brady and Yuille 4 develop an extremum principle that determines surface orientation from a 2-D contour. The principle maximises a compactness measure for the con-tour and they show that for irregular figures (or in the case of a sampled contour) their principle is roughly equivalent to Witkin's MLE. Their method though suf-fers from the same problems as Witkin's since it again involves a search in a 2-D space. In a recent paper 5 Blake develops a method for es-timating (a, T) as the orientation which makes the sec-ond moment tensor of the tangents to surface markings isotropic. Consider a tangent t ( on a point of a surface marking as an oriented line element which makes an an-gle 0i with one of the coordinate axes of the image plane. The second moment tensor T of these tangents can be written as: cos2

About this research paper

What this paper is about

ground Several researchers have investigated the role of texture as a basis for the recovery of surface orientation. Gib-son 1 was the first to address the problem of recovering the orientation of a plane covered with textural elements. He assumed that the density (number of elements per unit area) of these elements is uniform. Under this as-sumption he observed that the density gradient in the image specifies surface orientation. More recently 2, Witkin addresses the problem of recovering the slant a and tilt r of a planar surface viewed under orthographic projection. Let /? be the angle which a tangent at a point of a surface marking makes with the projection of a fixed coordinate axis of the image plane and a the angle the projection of that tangent makes with the coor-dinate axis. Under orthographic projection then Witkin develops a geometric model which relates a to /?: a — T + arctan /tan/A \\ COS <7 / Brady and Yuille 4 develop an extremum principle that determines surface orientation from a 2-D contour. The principle maximises a compactness measure for the con-tour and they show that for irregular figures (or in the case of a sampled contour) their principle is roughly equivalent to Witkin's MLE. Their method though suf-fers from the same problems as Witkin's since it again involves a search in a 2-D space. In a recent paper 5 Blake develops a method for es-timating (a, T) as the orientation which makes the sec-ond moment tensor of the tangents to surface markings isotropic. Consider a tangent t ( on a point of a surface marking as an oriented line element which makes an an-gle 0i with one of the coordinate axes of the image plane. The second moment tensor T of these tangents can be written as: cos2

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

ground Several researchers have investigated the role of texture as a basis for the recovery of surface orientation. Gib-son 1 was the first to address the problem of recovering the orientation of a plane covered with textural elements. He assumed that the density (number of elements per unit area) of these elements is uniform. Under this as-sumption he observed that the density gradient in the image specifies surface orientation. More recently 2, Witkin addresses the problem of recovering the slant a and tilt r of a planar surface viewed under orthographic projection. Let /? be the angle which a tangent at a point of a surface marking makes with the projection of a fixed coordinate axis of the image plane and a the angle the projection of that tangent makes with the coor-dinate axis. Under orthographic projection then Witkin develops a geometric model which relates a to /?: a — T + arctan /tan/A \\ COS <7 / Brady and Yuille 4 develop an extremum principle that determines surface orientation from a 2-D contour. The principle maximises a compactness measure for the con-tour and they show that for irregular figures (or in the case of a sampled contour) their principle is roughly equivalent to Witkin's MLE. Their method though suf-fers from the same problems as Witkin's since it again involves a search in a 2-D space. In a recent paper 5 Blake develops a method for es-timating (a, T) as the orientation which makes the sec-ond moment tensor of the tangents to surface markings isotropic. Consider a tangent t ( on a point of a surface marking as an oriented line element which makes an an-gle 0i with one of the coordinate axes of the image plane. The second moment tensor T of these tangents can be written as: cos2

Key concepts: Orthographic projection, Projection (relational algebra), Planar projection, Tangent, Orientation (vector space), Surface (topology), Projection plane, Tilt (camera)

Related papers

Back to paper searchBrowse research topicsOriginal source
Analysis of 3D Texture — Research Paper | ScholarLens