Granted Patent B2

US 12,518,066 · App. 16/951,776 · Granted Jan 6, 2026

Inventors: Emmanuel Gallo (Mougins, FR); Yan Fu (San Carlos, CA); Keith Alfaro (Brooklyn, NY); Manuel Martinez Alonso (Malaga, ES); Simranjit Singh Kohli (Cambridge, CA); Graceline Regala Amour (Somerville, MA)

Assignee: AUTODESK, INC.

G06F30/13G06F16/211G06F30/12G06N20/00

Abstract

Claims (85)

1 . A computer-implemented method for utilizing a synthetic floor plan drawing, comprising:

generating, in a computer, using generative design, a room layout for one or more rooms of the synthetic floor plan drawing, wherein the room layout comprises a room description that defines a semantic for one or more of the rooms;

obtaining, in the computer, symbol element information for a symbol element, wherein the symbol element information comprises an enumeration of room types in which the symbol element can appear and a location inside the one or more rooms in which the symbol element can appear;

programmatically generating, in the computer, synthetic symbol data comprising instances of the symbol element inside one or more of the rooms, wherein the instances populate the room layout based on the room layout and the symbol element information;

programmatically generating, in the computer, the synthetic floor plan drawing based on the synthetic symbol data and the room layout, wherein:

a ground truth of labels and locations of symbols in the synthetic floor plan drawing are known;

utilizing the synthetic floor plan drawing to train an object detection model, wherein the ground truth is utilized for orientation information learning;

utilizing the object detection model to autonomously recognize and extract symbol elements in a different floor plan drawing, wherein:

the object detection model segments the different floor plan drawing into multiple sections and identifies fixed patterns;

the object detection model selects one or more of the multiple sections that define a design drawing area; and

the object detection model recognizes the symbol elements in the selected one or more multiple sections;

generating a building information model (BIM) comprising the recognized and extracted symbol elements;

receiving user input relating to the BIM; and

refining the object detection model based on the user input.

2 . The computer-implemented method of claim 1 , wherein the room layout comprises:

a set of two-dimensional (2D) positions that correspond to a beginning and end of all walls.

3 . The computer-implemented method of claim 1 , wherein the generating instances of the symbol element further comprises:

generating a style for a dataset that can match a design of one or more designs, wherein for each of the one or more designs, a configuration file stores the symbol element information.

4 . The computer-implemented method of claim 1 , wherein:

the symbol element information is stored in a configuration file;

the configuration file comprises multiple blocks;

each of the multiple blocks is indexed with a unique identifier; and

each of the multiple blocks corresponds to a rendering of a different symbol element.

5 . The computer-implemented method of claim 4 , wherein:

each of the multiple blocks comprises the symbol element information;

the symbol element information comprises:

a scale of the symbol element;

the orientation of the symbol element;

a number of maximum instances;

a background property identifying whether the symbol element belongs to a background or a foreground; and

an associated label of the symbol element.

6 . The computer-implemented method of claim 1 , wherein:

the generating the synthetic floor plan drawing comprises randomly generating the synthetic floor plan drawing by calculating each instance randomly until the instance complies with the symbol element information.

7 . The computer-implemented method of claim 6 , further comprising:

determining that the instances of symbol elements in the synthetic floor plan drawing are unbalanced;

triangulating the synthetic floor plan drawing into one or more triangles; and

uniformly placing the symbol elements in the synthetic floor plan drawing using random uniform sampling inside the one or more triangles.

8 . The computer-implemented method of claim 7 , further comprising:

based on the determination that the instances of symbol elements are unbalanced, adjusting a weight of a symbol type, wherein the weight affects how many instances of the symbol type are placed into the synthetic floor plan drawing.

9 . The computer-implemented method of claim 1 , further comprising:

adding random additional information to the synthetic floor plan drawing to improve a machine learning model.

10 . A computer-implemented system for generating a synthetic floor plan drawing, comprising:

(a) a computer having a memory;

(b) a processor executing on the computer;

(c) the memory storing a set of instructions, wherein the set of instructions, when executed by the processor cause the processor to perform operations comprising:

(i) generating, using generative design, a room layout for one or more rooms of the synthetic floor plan drawing, wherein the room layout comprises a room description that defines a semantic for one or more of the rooms;

(ii) obtaining symbol element information for a symbol element, wherein the symbol element information comprises an enumeration room types in which the symbol element can appear and a location inside the one or more rooms in which the symbol element can appear;

(iii) programmatically generating synthetic symbol data comprising instances of the symbol element inside one or more of the rooms, wherein the instances populate the room layout based on the room layout and the symbol element information;

(iv) programmatically generating synthetic floor plan drawing based on the synthetic symbol data and the room layout, wherein:

(A) a ground truth of labels and locations of symbols in the synthetic floor plan drawing are known;

(v) utilizing the synthetic floor plan drawing to train an object detection model, wherein the ground truth is utilized for orientation information learning;

(vi) utilizing the object detection model to autonomously recognize and extract symbol elements in a different floor plan drawing, wherein:

(A) the object detection model segments the different floor plan drawing into multiple sections and identifies fixed patterns;

(B) the object detection model selects one or more of the multiple sections that define a design drawing area; and

(C) the object detection model recognizes the symbol elements in the selected one or more multiple sections;

(vii) generating a building information model (BIM) comprising the recognized (vii) and extracted symbol elements;

(viii) receiving user input relating to the BIM; and

(ix) refining the object detection model based on the user input.

11 . The computer-implemented system of claim 10 , wherein the room layout comprises:

a set of two-dimensional (2D) positions that correspond to a beginning and end of all walls.

12 . The computer-implemented system of claim 10 , wherein the generating instances of the symbol element further comprises:

generating a style for a dataset that can match a design of one or more designs, wherein for each of the one or more designs, a configuration file stores the symbol element information.

13 . The computer-implemented system of claim 10 , wherein:

the symbol element information is stored in a configuration file;

the configuration file comprises multiple blocks;

each of the multiple blocks is indexed with a unique identifier; and

each of the multiple blocks corresponds to a rendering of a different symbol element.

14 . The computer-implemented system of claim 13 , wherein:

each of the multiple blocks comprises the symbol element information;

the symbol element information comprises:

a scale of the symbol element;

the orientation of the symbol element;

a number of maximum instances;

a background property identifying whether the symbol element belongs to a background or a foreground; and

an associated label of the symbol element.

15 . The computer-implemented system of claim 10 , wherein:

the generating the synthetic floor plan drawing comprises randomly generating the synthetic floor plan drawing by calculating each instance randomly until the instance complies with the symbol element information.

16 . The computer-implemented system of claim 15 , wherein the operations further comprise:

determining that the instances of symbol elements in the synthetic floor plan drawing are unbalanced;

triangulating the floor plan drawing into one or more triangles; and

uniformly placing the symbol elements in the synthetic floor plan drawing using random uniform sampling inside the one or more triangles.

17 . The computer-implemented system of claim 16 , wherein the operations further comprise:

based on the determination that the instances of symbol elements are unbalanced, adjusting a weight of a symbol type, wherein the weight affects how many instances of the symbol type are placed into the synthetic floor plan drawing.

18 . The computer-implemented system of claim 10 , wherein the operations further comprise:

adding random additional information to the synthetic floor plan drawing to improve a machine learning model.

Assignments (1)

Continuity (3)

References Cited (63)