Can machines interpret line drawings

Journal of Graphic Science of Japan, 2003

In engineering design, much time and effort could be saved if a computer could interpret an engineer's initial concept drawings. This paper describes a prototype system for interpreting line drawings as solid polyhedral objects. It concentrates on four stages of processing: line labelling, identification of parallel lines, inflation to 2 1/2 D, and determination of hidden topology. Although the system is often successful, particularly with simple drawings, there are also cases where the correct interpretation is obvious to a human but cannot be determined using existing algorithms. The paper gives examples both of the successes of the approach and of typical cases where it cannot replicate human skill.

1987

Reviewed by WALTER WHITELEY The author obtained his advanced training in the Engineering Division of the University of Tokyo and is currently an Associate Professor in the Department of Mathematical Engineering and Instrumentation Physics at that university. His primary research interests involve the topics of the book under review and related matroid theory and combinatorial algorithms.

Communications of the ACM, 1992

IEEE Transactions on Pattern Analysis and Machine Intelligence, 1984

The two parts, which this paper is composed of, deal each with scene interpretation via gaining understanding of the faces of the objects in the scene. The first part extends the set of rules defined in a previous work regarding the assembling of all lines belonging to the same face. The set of rules, originally defined for curved object, can be extended if we confine ourselves to polyhedra. In the second part, a new concept is defined and developed, which leads to a new way of looking at polyhedral line drawings. It puts under the same roof almost all consistency checks known for polyhedra, in a natural and simple way. Geometric inconsistencies as well as interpretations inconsistencies are treated uniformly and in a straightforward manner. Through this concept a way is suggested for acquiring some understanding of back faces, and for suggesting plausible interpretation for them. The generality of this concept is demonstrated through the fact that previously known catalogs of labeled junctions can be derived directly from this concept.

We describe a two-stage approach for interpreting line drawings of curved objects. In the first stage, the user en- ters a natu-ral line drawing of a polyhedral template; this is automatically interpreted as the corresponding poly- hedral object. In the second stage, the user enters freehand curves; by relating these to the template, a curved ob- ject can be constructed

IEEE Transactions on Pattern Analysis and Machine Intelligence, 1985

This paper deals with the interpretation and feasibility check of line drawings representing polyhedral scenes. The polyhedra are of general types and there are no restrictions on camera position. The geometric consistency check and the line labeling are carried out through constructions in the image plane. An algorithm for the geometric construction is suggested, and the necessary conditions for these constructions are discussed. The image plane construction can be used for preparing labeled junictioii catalogs for junctions other than trihedral. In addition the paper analyzes the relation between the image plane construction and the gradient space construction suggested by Mackworth [7] for the same purpose. Index Terms-Computer vision, machine intelligence, polyhedral line drawing, projective geometry, scene analysis.

2004

As part of the goal of automatic creation of B-rep models of engineering objects from freehand sketches, we seek to take a single line drawing (with hidden lines removed), and from it deduce an initial 3D geometric realisation of the visible part of the drawn object. Junction and line labels, and provisional depth coordinates, are key parts of this frontal geometry.

2002

International Journal on Document Analysis and Recognition, 1999

Converting paper-based engineering drawings into CAD model files is a tedious process. Therefore, automating the conversion of such drawings represents tremendous time and labor savings. We present a complete system which interprets such 2D paper-based engineering drawings, and outputs 3D models that can be displayed as wireframes. The system performs the detection of dimension sets, the extraction of object lines, and the assembly of 3D objects from the extracted object lines. A knowledge-based method is used to remove dimension sets and text from ANSI engineering drawings, a graphics recognition procedure is used to extract complete object lines, and an evidential rulebased method is utilized to identify view relationships. While these methods are the subject of several of our previous papers, this paper focuses on the 3D interpretation of the object. This is accomplished using a technique based on evidential reasoning and a wide range of rules and heuristics. The system is limited to the interpretation of objects composed of planar, spherical, and cylindrical surfaces. Experimental results are presented.

2005

Freehand sketching is an important part of the conceptual design process, and the increasing number of recent sketching applications shows a developing awareness of this importance. We aim to provide an automated tool to turn engineers� freehand sketches into CAD models. This would allow engineers to spend their time more productively and to be more creative. One natural component of such a tool would be a process for identifying axially-aligned planes implied by a natural line drawing. We present an algorithm for identifying such planes. We illustrate its utility by presenting two uses: identifying planes of mirror symmetry in objects and constructing the hidden topology of objects.