Method and apparatus for image assisted modeling of three-dimensional scenes

• US 7,193,633 B1
• Filed: 04/27/2000
• Issued: 03/20/2007
• Est. Priority Date: 04/27/2000
• Status: Expired due to Fees

• Current Assignee: Adobe Systems Incorporated (Adobe Inc.)
• Original Assignee: Adobe Systems Incorporated (Adobe Inc.)
• Inventors: Seidl, Robert, Reinhardt, Tilman
• Primary Examiner(s): Tung; Kee M.
• Assistant Examiner(s): HARRISON, CHANTE E
• Application Number:
  US09/562,212
• Time in Patent Office: 2,518 Days
• Field of Search: 345/419, 345/420, 345/429, 345/427, 345/619, 345632-633, 345/627, 345/629, 345/582, 345/620
• US Class Current: 345/629
• CPC Class Codes: G06T 17/10 Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes

Abstract

A computer-assisted technique for constructing a three-dimensional model on top of one or more images (e.g., photographs) such that the model'"'"'s parameters automatically match those of the real world object depicted in the photograph(s). Camera parameters such as focal length, position, and orientation in space may be determined from the images such that the projection of a three-dimensional model through the calculated camera parameters matches the projection of the real world object through the camera onto the image surface. Modeling is accomplished using primitives, such as boxes or pyramids, which may be intuitively manipulated to construct the three-dimensional model on a video display or other display screen of a computer system with a two-dimensional input controller (e.g., a mouse, joystick, etc.) such that the displayed three-dimensional object manipulation emulates physical three-dimensional object manipulation. Camera and primitive parameters are incrementally updated to provide visual feedback of the effect of additional constraints on the three-dimensional model, making apparent which user action may have been responsible for any failure to provide a modeling solution and, thus, allowing for rapid reversal and correction thereof. Surface properties (i.e., textures) may be extracted from the images for use in the three-dimensional model.

Claims

1. A computer-assisted method, comprising:

constructing a three-dimensional model on top of one or more digital images through direct manipulation via a cursor-control device of one or more renderings of predefined geometric primitives selected to approximate items depicted in the one or more digital images and which (i) collectively make up the three-dimensional model, and (ii) are super-imposed on the one or more digital images according to estimated virtual camera parameters determined from the one or more digital images, wherein the estimated virtual camera parameters include position, orientation, and focal length of the virtual camera, and wherein the estimated virtual camera parameters are automatically and continually updated during the manipulation of the predefined geometric primitives so as to accurately project the primitives in accordance with the manipulation;

constraining the primitives in relation to the items in the one or more digital images such that projections of the primitives in accordance with the estimated virtual camera parameters over corresponding ones of the items in the one or more digital images yields two-dimensional projected geometries matching those of the corresponding items in the one or more digital images so that parameters of the three-dimensional model automatically match corresponding parameters of objects depicted in the one or more images; and

automatically extracting surface properties from the one or more digital images for use as textures for the geometric primitives in the three-dimensional model and;

displaying the three-dimensional model.

2. The computer-assisted method of claim 1 wherein the one or more digital images are derived from one or more photographs.