IP Library › Granted Patent US 12,511,804
Granted Patent B2
US 12,511,804 · App. 18/216,900 · Granted Dec 30, 2025

Systems and methods for automating conversion of drawings to indoor maps and plans

Inventors: Ege Çetintaş (Ankara, TR); Melih Peker (Ankara, TR); Can Tunca (Istanbul, TR)
Assignee: Pointr Limited
G06T11/60G06F30/13G06N20/00G06T11/20
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,511,804
App. No.
18/216,900
Granted
Dec 30, 2025
Kind
B2
Abstract

Automating conversion of drawings to indoor maps and plans. One example is a computer-implemented method comprising: preprocessing a CAD drawing to create a text database containing text from the CAD drawing and associations of the text with locations within the CAD drawing; determining a floor depicted in the CAD drawing, the determining results in a floor-level outline; identifying a plurality of room-level outlines within the floor-level outline, the plurality of room-level outlines corresponds to a respective plurality of rooms; selecting a name of a first room from the plurality of rooms, the selecting based on text within the text database; and creating an indoor map including the name of the first room, the name of the first room associated with a location of the first room within the floor-level outline.

Claims (56)

1 . A computer-implemented method of extracting room names from CAD drawings, the method comprising:

preprocessing, by a device, a CAD drawing to create a text database containing text from the CAD drawing and associations of the text with locations within the CAD drawing;

determining, by a device, a floor depicted in the CAD drawing, the determining results in a floor-level outline;

identifying, by a device, a plurality of room-level outlines within the floor-level outline, the plurality of room-level outlines corresponds to a respective plurality of rooms;

selecting, by a device, a name of a first room from the plurality of rooms, the selecting based on text within the text database, wherein selecting the name for the first room further comprises:

grouping text in the text database associated with the first room, the grouping creates a first grouped text,

lemmatizing words of the first grouped text to form a first plurality of lemmas;

assigning a member probability to each lemma of the first plurality of lemmas, each member probability representing likelihood of a lemma being a member of the name of the first room,

generating combinations of the first plurality of lemmas having probability above a predetermined threshold, the generating creates a first plurality of monikers,

determining that the first plurality of monikers each lack a lemma having member probability above a predetermined threshold,

retraining a machine-learning model configured to generate the combinations of the first plurality of lemmas,

assigning a moniker probability to each moniker of the first plurality of monikers, each moniker probability representing likelihood of a moniker being the name of the first room, and

selecting the name of the first room based on the first plurality of monikers and the moniker probabilities; and

creating, by a device, an indoor map including the name of the first room, the name of the first room associated with a location of the first room within the floor-level outline.

2 . The computer-implemented method of claim 1 further comprising, after grouping but before lemmatizing, separating words and numbers of the first grouped text.

3 . The computer-implemented method of claim 1 further comprising, after selecting the name of the first room, de-lemmatizing the name of the first room.

4 . The computer-implemented method of claim 1 further comprising selecting a name for a second room, distinct from the first room, the selecting by:

grouping text in the text database associated with the second room, the grouping creates a second grouped text;

lemmatizing words of the second grouped text to form a second plurality of lemmas;

assigning a member probability to each lemma of the second plurality of lemmas, each member probability of the second plurality of lemmas representing likelihood of a lemma being a member of the name of the second room;

generating combinations of the second plurality of lemmas having probability above a predetermined threshold, the generating creates a second plurality of monikers;

assigning a moniker probability to each moniker of the second plurality of monikers, each moniker probability of the second plurality of monikers representing likelihood of a moniker being the name of the second room; and

selecting the name of the second room based on the second plurality of monikers and the moniker probabilities of the second plurality of monikers.

5 . The computer-implemented method of claim 4 further comprising, after selecting the name of the second room, de-lemmatizing the name of the second room.

6 . The computer-implemented method of claim 1 further comprising, prior to grouping text, removing predetermined characters.

7 . The computer-implemented method of claim 6 wherein removing predetermined characters further comprises removing characters associated with formatting of the text.

8 . The computer-implemented method of claim 1 wherein determining the floor depicted in the CAD drawing further comprises receiving the floor-level outline from a user.

9 . A computer system comprising:

a processor;

a memory coupled to the processor, the memory storing instructions that, when executed by the processor, cause the processor to:

preprocess a CAD drawing to create a text database containing text from the CAD drawing and associations of the text with locations within the CAD drawing;

determine a floor depicted in the CAD drawing, the determining results in a floor-level outline;

identify a plurality of room-level outlines within the floor-level outline, the plurality of room-level outlines corresponds to a respective plurality of rooms;

select a name of a first room from the plurality of rooms, the selection based on text within the text database, wherein when the processor selects the name for the first room, the instructions further cause the processor to:

group text in the text database associated with the first room, the grouping creates a first grouped text,

lemmatize words of the first grouped text to form a first plurality of lemmas,

assign a member probability to each lemma of the first plurality of lemmas, each member probability representing likelihood of a lemma being a member of the name of the first room,

generate combinations of the first plurality of lemmas having probability above a predetermined threshold, the generating creates a first plurality of monikers,

determine that the first plurality of monikers each lack a lemma having member probability above a predetermined threshold,

trigger retraining of a machine-learning model configured to generate the combinations of the first plurality of lemmas,

assign a moniker probability to each moniker of the first plurality of monikers, each moniker probability representing likelihood of a moniker being the name of the first room, and

select the name of the first room based on the first plurality of monikers and the moniker probabilities; and

create an indoor map including the name of the first room, the name of the first room associated with a location of the first room within the floor-level outline.

10 . The computer system of claim 9 wherein the instructions further cause the processor to, after grouping but before lemmatizing, separate words and numbers of the first grouped text.

11 . The computer system of claim 10 wherein the instructions further cause the processor to, after select the name of the first room, de-lemmatize the name of the first room.

12 . The computer system of claim 9 wherein the instructions further cause the processor to select a name for a second room, distinct from the first room, by causing the processor to:

group text in the text database associated with the second room, the grouping creates a second grouped text;

lemmatize words of the second grouped text to form a second plurality of lemmas;

assign a member probability to each lemma of the second plurality of lemmas, each member probability of the second plurality of lemmas representing likelihood of a lemma being a member of the name of the second room;

generate combinations of the first plurality of lemmas having probability above a predetermined threshold, the generating creates a second plurality of monikers;

assign a moniker probability to each moniker of the second plurality of monikers, each moniker probability of the second plurality of monikers representing likelihood of a moniker being the name of the second room; and

select the name of the second room based on the first plurality of monikers and the moniker probabilities of the second plurality of monikers.

13 . The computer system of claim 12 wherein when the instructions further cause the processor to, after selecting the name of the second room, de-lemmatize the name of the second room.

14 . The computer system of claim 9 wherein the instructions further cause the processor to, prior to grouping text, remove predetermined characters.

15 . The computer system of claim 14 wherein when the processor removes predetermined characters, the instructions cause the processor to remove characters associated with formatting of the text.

16 . The computer system of claim 9 wherein when the processor determines the floor depicted in the CAD drawing, the instructions further cause the processor to receiving the floor-level outline from a user.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 7, 2023
From: ÇETINTAS, EGE; PEKER, MELIH; TUNCA, CAN
To: POINTR LIMITED
Reel/Frame 064235/0179 →
Continuity (4)
Continuation In Part 18052852 · Nov 4, 2022
Continuation 17732652 · Apr 29, 2022
Provisional Application 63318522 · Mar 10, 2022
Related Publication 20230377227A1 · Nov 23, 2023
References Cited (38)
US 7227893B1 · Srinivasa et al. · 2007 [cited by applicant]
US 10539424B2 · Roy et al. · 2020 [cited by applicant]
US 11156465B2 · Roy et al. · 2021 [cited by applicant]
US 11190902B1 · Brosowsky et al. · 2021 [cited by applicant]
US 11514633B1 · Cetintas et al. · 2022 [cited by applicant]
US 11657555B1 · Cetintas et al. · 2023 [cited by applicant]
US 11725948B2 · Roy et al. · 2023 [cited by applicant]
US 11769287B1 · Cetintas et al. · 2023 [cited by applicant]
US 20040049307A1 · Beatty et al. · 2004 [cited by applicant]
US 20100214290A1 · Shiell et al. · 2010 [cited by applicant]
US 20150199557A1 · Zhang et al. · 2015 [cited by applicant]
US 20190311533A1 · Doh · 2019 [cited by examiner]
US 20200126003A1 · McLaughlin, III · 2020 [cited by examiner]
US 20210073302A1 · Srinivasaraghavan · 2021 [cited by examiner]
US 20210073433A1 · Austern · 2021 [cited by examiner]
US 20210073435A1 · Segev et al. · 2021 [cited by applicant]
US 20210150088A1 · Gallo et al. · 2021 [cited by applicant]
US 20210241072A1 · Katzenelson · 2021 [cited by examiner]
US 20210409903A1 · Shapiro et al. · 2021 [cited by applicant]
US 20220035973A1 · Liebman et al. · 2022 [cited by applicant]
US 20220138621A1 · Patil et al. · 2022 [cited by applicant]
US 20220147026A1 · Poelman et al. · 2022 [cited by applicant]
US 20230157506A1 · Tamino et al. · 2023 [cited by applicant]
US 20230306664A1 · Cetintas et al. · 2023 [cited by applicant]
CN 111854758A · 2020 [cited by applicant]
Extended European Search Report mailed Aug. 3, 2023 for European Application No. 23155925.3-1009, 6 pages. [cited by applicant]
Simonsen, Christoffer P. et al., “Generalizing Floor Plans Using Graph Neural Networks”, International Conference on Image Processing (ICIP), Sep. 2021, pp. 654-658. [cited by applicant]
“Import AutoCAD Files in 1 Minute Only”, https://www.mapwize.io/news/2019-01-import-autocad-files-in-1-minute--only/, ServiceNow, Jan. 2019, 2 pages. [cited by applicant]
“Manage all your indoor maps from a dedicated platform”, https://visioglobe.com/indoor-mapping-tools, Visioglobe.com, Accessed: Jul. 6, 2022, 6 pages. [cited by applicant]
“Integrate Indoor Mapping Data Format {IMDF) Using FME”, https://visioglobe.com/indoor-mapping-tools, Safe Software Accessed: Jul. 6, 2022, 9 pages. [cited by applicant]
DIY Augmented Reality Indoor Mapping for your venue, https://www.safe.com/blog/2018/09/diy-augmented-reality-indoor-mapping-for-your-venue/, Safe Software, Accessed: Jul. 6, 2022, 11 pages. [cited by applicant]
“Covert DWG to IMDF,” https://www.safe.com/blog/2018/09/diy-augmented-reality-indoor-mapping-for-your-venue/, Safe Software, Accessed: Jul. 6, 2022, 6 pages. [cited by applicant]
Sarker, Iqbal H., “Deep Learning: A Comprehensive Overview of Techniques, Taxonomy, Applications and Research Directions”, SN Computer Science, Aug. 18, 2021, 20 pages. [cited by applicant]
Huang H.C., et al., “Graph Theory-based Approach for Automatic Recognition of CAD Data”, ScienceDirect, Engineering Applications of Artificial Intelligence 21, 2008, 7 pages. [cited by applicant]
Tang, Rui, et al., “Automatic Structural Scene Digitalization”, Centre for the Analysis of Motion, Entertainment Research and Applications, University of Bath UK, Nov. 17, 2017, 16 pages. [cited by applicant]
Tang, Hao, et al., “Automatic Pre-Journey Indoor Map Generation Using AutoCAD Floor Plan”, The Journal on Technology and Persons with Disabilities, Sep. 2017, 17 pages. [cited by applicant]
“Apple Indoor Maps and Positioning”, Apple Inc., Apr. 2021, 22 pages. [cited by applicant]
De Las Heras, Lluis-Pere, et al., “Statistical Segmentation and Structural Recognition for Floor Plan Interpretation”, German Research Center for AI, Dec. 2013, 18 pages. [cited by applicant]