[go: up one dir, main page]

WO2025099621A1 - Procédé semi-automatique de recherche d'images d'ouvrages de maçonnerie - Google Patents

Procédé semi-automatique de recherche d'images d'ouvrages de maçonnerie Download PDF

Info

Publication number
WO2025099621A1
WO2025099621A1 PCT/IB2024/061030 IB2024061030W WO2025099621A1 WO 2025099621 A1 WO2025099621 A1 WO 2025099621A1 IB 2024061030 W IB2024061030 W IB 2024061030W WO 2025099621 A1 WO2025099621 A1 WO 2025099621A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
user
target
images
digital
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/IB2024/061030
Other languages
English (en)
Inventor
Rossella CORRAO
Calogero VINCI
Erica LA PLACA
Nicola BARBUTI
Tommaso CALDAROLA
Enrico GENOVA
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Universita degli Studi di Palermo
Original Assignee
Universita degli Studi di Palermo
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Universita degli Studi di Palermo filed Critical Universita degli Studi di Palermo
Publication of WO2025099621A1 publication Critical patent/WO2025099621A1/fr
Pending legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/53—Querying
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/53—Querying
    • G06F16/532—Query formulation, e.g. graphical querying
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5854—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using shape and object relationship

Definitions

  • the present invention relates to a semi-automatic method for searching for masonry work images.
  • the object of the present invention is to propose a particularly reliable and efficient method for searching through masonry work images that is quick and easy to utilize on the part of a user.
  • Another object of the invention is to meet the aforementioned need to obtain, quickly and easily for the user, information relating to a masonry work of interest.
  • FIG. 1 is a flow chart of the semi- automatic method of searching for masonry work images according to the invention.
  • Figure 2 is an example of a target image equipped with markers for connecting to digital contents.
  • the semi- automatic masonry work image search method provides for making available a collection 10 of target digital masonry work images. At least a part of such target images is associated with at least one further digital content 12.
  • the digital target image collection 10 is made accessible by a processing device of a user.
  • the user selects a digital source image 14 (step 50 of the flow diagram in Figure 1), representing at least one portion of a work of masonry.
  • the digital source image 14 is for example extracted from the same collection 10 of digital target images.
  • the user extracts a region of interest 16 (ROT, step 60) from such masonry work portion.
  • region of interest 16 represents a stone element, a brick or an ashlar, or a set of such masonry work elements.
  • step 70 which associates at least one search parameter (SP) and at least one configuration parameter
  • one or more target images 18 are identified within the collection of digital target images (step 80).
  • the comparison with the source model is performed using digital target images provided with further digital content 12.
  • Target images 18 are identified if they include a target portion having a shape that is similar to the shape of the region of interest.
  • the user is then provided with the identified target image (s) 18, by means of the user interface module, together with the respective further digital contents 12.
  • the extraction of the region of interest 16 occurs by selecting those pixels wherein the contrast thereof, with respect to neighboring pixels, exceeds a pre-established threshold.
  • the creation of the source model further comprises editing, on the part of the user by means of the user interface module, one or more features of the image of the region of interest (step
  • such features comprise: angle, rotation, scale, contrast, brightness, color, focus, distortion, deformation.
  • the further digital contents 12 associated with the target images 18 of the collection of target images are linked to respective portions of the target images by means of a graphically highlighted marker ("markers" 20) on the target image 18 (see Figure 2).
  • the digital source image 14' is obtained by scanning an actual masonry work 22 on the part of the user by means of a digital camera 24.
  • the user may immediately access the digital contents 12 associated with the identified target image 18 (step 90).
  • an image insertion function (step 100) is initiated, according to which the user sends, by means of the user interface module, the scanned image 14' to a system administrator (i.e., the party who also administers the collection (s) of digital masonry work images 10) for the insertion of such image into the collection 10.
  • a system administrator i.e., the party who also administers the collection (s) of digital masonry work images 10.
  • the image insertion function furthermore comprises a description interface that is suitable for allowing the user to associate relevant information with the scanned image.
  • the user may interact with digital image collections relating to masonry of different historical periods, using a user interaction method that is based upon a process based upon the following steps:
  • the method is based upon the existence, or creation, of a collection 10 populated with images of masonry and related descriptions 12; it may be useful to take significant samples of each type of artifact such as to cluster the same or similar artifacts.
  • ROI graphical regions 16
  • ROI(s) use the ROI(s) as search key(s) through which to recover graphically similar regions within one or more target images (digital target images), included within one or multiple digital collections 10, existing on the local network or on the internet.
  • the user generates the ROI by pointing to a segment within the selected image.
  • the algorithm activates an automatic model generation function by detecting contrast factors between pixels on the image layout, defining them in parameters and relative thresholds that the user may modify by means of interaction functions.
  • the user may optimize the search in terms of performance and quality of the results by setting and customizing parameters and relative thresholds (step 55 in Figure 1) by virtue of the availability of the following operators:
  • the settings may be managed using, as parameters and relative base thresholds for searching, both a complete set of preset primitives 30 and graphical functions 32 that may be used on the part of the user to customize the parameters by acting on the related thresholds.
  • Some of the main graphical functions provided by the algorithm are: angle, rotation, scale, contrast, brightness, color, focus, distortion, deformation.
  • the search is performed by means of a user interface module, provided with the following features: [0042]- real-time connection to locally available digital collections, using a selection feature of the repository (ies) of interest; it is also possible to connect the interface to collections available on the internet;
  • ROI Region of Interest
  • the matching algorithm analyzes the images of the collection (s) according to the settings of the user and retrieves, within the target images, those ROIs that it considers to be similar to that of the source model.
  • the graphic matching function recognizes those ROIs within images that are similar to each other based upon the shape thereof and does not require preliminary segmentation of the pages of the document.
  • ROIs are defined as model(s) (see paragraph A.3 below), the algorithm finds similarities 18 thereof, if any, within the collections of target images.
  • the user may generate n ROI/models within n images and parallelize searches for multiple models thereby optimizing collection analysis and similarity recovery times.
  • the user may obtain the following information for each recovered model:
  • model ROI characterizes and defines the internal representation of the part of the image to be used as the search element.
  • the source image format for defining the model may be one of the common digital formats, such as TIFF, BMP, GIF, JPEG, PPM, PGM, PNG, PBM, etc.
  • the image may be of any shape (elliptical, circular, polygonal, or even freehand outlined) and may have an arbitrary angle.
  • the optimization of the search process starts from the definition of a good model.
  • the ROT to be used as a search key may be modeled and extracted automatically: the user points to one or more segments of interest within the image chosen on the part of the user in the dataset; the algorithm responds to the pointing action by automatically generating the ROT to be used as a model.
  • the algorithm generates the model by selecting within the region all those pixels the wherein the contrast thereof, with respect to neighboring pixels, exceeds a certain threshold: typically, in the presence of contrast factors within the pointed to ROT, the algorithm selects and uses all of the pixels that belong to the contour of an element present within the region.
  • the algorithm provides the user with functions for improving the quality of the model by intervening upon the pixels in order to eliminate noise that may disturb the matching.
  • each template created may be stored and saved in order to retrieve and edit it for further use.
  • the parameters of the contrast thresholds must be chosen by including within the model all of those pixels that are significant for the identification of the object.
  • the term "Significant pixels” refers to those pixels that characterize the object and allow the searched for shape to be clearly differentiated from other objects and from the background.
  • thresholds are typically preset that automatically identify the ashlar and/or brick as the minimum unit of the architectures and reduce noise factors that are inconsistent with the correct contrast factors.
  • the model must have a minimum noise level and exclude regions that are not relevant to the segment to be searched.
  • [0062]- number of pyramid levels constituting the model the set of images at different resolutions representing the same source image, ordered by decreasing resolution, is the pyramid, and the images within a pyramid constitute the pyramid levels, wherein the vertex of the pyramid is the image at the lowest resolution.
  • This is an important factor for the performance and precision of the result: as a general rule, a good result is obtained if a region of interest with a width of 2 LevelNumber - 1 pixel, i.e. 8 pixels wide, allows four pyramid levels to be used. After setting the region, the image may be used as a model for creating the reference shape;
  • the model may use the following basic configuration parameters:
  • the model may also utilize the following advanced configuration parameters: [0069]- completeness: determines the trade-off between the efficiency and effectiveness of the search results. A low value determines a complete, but rather slow search; the higher the value, the faster the search, but the completeness suffers (i.e., an occurrence of the model may not be found even if it is visible within the image);
  • [0070]- sub-pixels defines the accuracy by selecting the precision factor in the calculation of the position, orientation, and scale; the accuracy of the estimated scale and orientation depends upon the size of the object: the larger the size, the more precise the orientation and scale will be;
  • [0071]- deformation if there are target images wherein searched objects are reproduced with slight deformation with respect to the model ROI, the user may use a deformation parameter that expresses how many deviation pixels must be tolerated between the contours found in the target image and those of the model. This parameter should be set to the lowest possible value; high values may only be used for targeted searches. In fact, the greater this value, the greater the risk of recovering 'false positives' and, at the same time, the greater the processing time.
  • This step (110) involves the analysis and selection of images included within one or more collections allocated to local storage or the cloud that are to be integrated with information by means of the association of digital extensions.
  • the analysis and selection is carried out by the administrator of the collection (s) 10 and allows for the identification of: [0073]- type and formats of the digital content to be associated as extensions: the content to be associated may be of any format and type; links may also be associated with specially structured interfaces for activating user profiling procedures, setting up descriptive cards, and entering data that is of interest to both the consumer user and the operator user;
  • the markup is performed by the administrator user by means of selectively pointing to those segments to be connected to the extensions. Following the selective pointing, the algorithm automatically determines those pixels that comprise the segment (s) to be marked.
  • the marked segment may be highlighted using:
  • the graphic may be customized by the administrator user; for example, it may be decided to use a stone or ashlar element as an access marker 20, wherein the user view will appear in a different color with respect to the rest of the layout ( Figure 2).
  • the administrator user utilizes call to action features in order to connect digital contents allocated to local storage or the internet.
  • the user selects contents and associates them by means of automatic import functions of files of various types or external links, which may be masked using html code (step 130).
  • the administrator user may associate artificial intelligence platforms and use prompts in order to generate responsive interactions wherefrom contents are created in real time.
  • the user interaction is achieved by reusing and customizing some specific features of a third-party application that may be used on mobile devices.
  • the application automatically connects to digital collections of architecture images that are allocated locally or in the cloud. By downloading it onto the mobile device thereof, the user automatically has available those digital collections that are connected thereto.
  • the features of the application allows the user to interact on site with original architecture by means of scanning masonry artifacts 22 using the camera 24 of the mobile device thereof.
  • the algorithm automatically retrieves it and displays it on the mobile device screen (step 90).
  • the algorithm responds negatively to the search and automatically activates an interaction function based upon which the user may:
  • [0086]- send the scanned image to the system administrator for inclusion within the collection by associating relevant information entered into a description interface (step 100);
  • the user may scan the original architecture from different positions several times and send one, some, or all of the images to the administrator; the administrator decides whether to keep them all or only part of them, and selects the image to be used as a target reference for matching.
  • the user interaction differs also in the use of target images on the screen of mobile devices.
  • the target image 18 is already provided with digital extensions 12, these are highlighted with markers 20 on the image displayed on the screen, and the user may interact with them and explore them through the Tangible User Interface (TUI) feature (step 140); the user may also use an input function to propose further extensions to be associated by sending the information to the administrator.
  • TTI Tangible User Interface
  • the user may use the input function to propose extensions to be associated by sending the information to the administrator (step 100); by means of a profiling log, if an experienced user, they may access a dedicated interface wherein to enter and send scientific information; if a non-expert user, they may access a text interface wherein to enter and send generic information.
  • the administrator evaluates the information and decides whether to associate it in whole or in part with the target image.
  • the present invention also relates to a computer product, i.e., software, comprising portions of code which, when executed on a processing device, such as a smartphone or tablet, cause the processing device to perform the semi-automatic image search method described above.
  • a computer product i.e., software
  • a processing device such as a smartphone or tablet
  • the present invention also relates to a user processing device, such as a computer, a tablet, or a smartphone, configured with an application to access a database containing a collection of target digital masonry work images, wherein each target image is associated with at least one further piece of digital content, and to perform the steps of the semi-automatic image search method described above.
  • a user processing device such as a computer, a tablet, or a smartphone, configured with an application to access a database containing a collection of target digital masonry work images, wherein each target image is associated with at least one further piece of digital content, and to perform the steps of the semi-automatic image search method described above.
  • the subject matter of the present invention also encompasses an administrator processing device configured to generate a collection of digital target masonry work images, wherein each target image is associated with at least one further piece of digital content, and to receive data from the user processing device.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

La présente invention concerne un procédé semi-automatique de recherche d'images d'ouvrages de maçonnerie, comprenant : - la recherche dans une ou plusieurs collections d'images numériques relatives à des architectures de maçonnerie sur la base de caractéristiques de motif graphique qui agissent sur la disposition des images, à l'aide de paramètres qui peuvent également être gérés depuis le côté utilisateur; - l'association d'extensions, ou contenus, numériques, à des collections d'images numériques relatives à des architectures de maçonnerie; - l'expérience utilisateur interactive d'architectures de maçonnerie en temps réel : en utilisant les caractéristiques d'une application téléchargée sur un téléphone intelligent/tablette, l'utilisateur effectue un balayage en présence de la maçonnerie; s'il existe une image correspondante dans la collection connectée à l'application, et que l'image a été enrichie à l'aide d'extensions numériques, l'utilisateur peut interagir en temps réel et en présence de l'image sur le téléphone intelligent/la tablette de celui-ci et utiliser ces extensions numériques qui fournissent des informations pertinentes.
PCT/IB2024/061030 2023-11-10 2024-11-07 Procédé semi-automatique de recherche d'images d'ouvrages de maçonnerie Pending WO2025099621A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
IT102023000023781 2023-11-10
IT202300023781 2023-11-10

Publications (1)

Publication Number Publication Date
WO2025099621A1 true WO2025099621A1 (fr) 2025-05-15

Family

ID=89897306

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2024/061030 Pending WO2025099621A1 (fr) 2023-11-10 2024-11-07 Procédé semi-automatique de recherche d'images d'ouvrages de maçonnerie

Country Status (1)

Country Link
WO (1) WO2025099621A1 (fr)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120095982A1 (en) * 2000-11-13 2012-04-19 Lennington John W Digital Media Recognition Apparatus and Methods
US20180165540A1 (en) * 2011-12-08 2018-06-14 Excalibur Ip, Llc Image object retrieval

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120095982A1 (en) * 2000-11-13 2012-04-19 Lennington John W Digital Media Recognition Apparatus and Methods
US20180165540A1 (en) * 2011-12-08 2018-06-14 Excalibur Ip, Llc Image object retrieval

Similar Documents

Publication Publication Date Title
JP3469345B2 (ja) 画像のファイリング装置及びファイリング方法
CN101453575B (zh) 一种视频字幕信息提取方法
CN110196917B (zh) 个性化logo版式定制方法、系统和存储介质
CN114092938A (zh) 图像的识别处理方法、装置、电子设备及存储介质
JP4757001B2 (ja) 画像処理装置、画像処理方法
CN112926601A (zh) 基于深度学习的图像识别方法、装置、设备及存储介质
CN110992384A (zh) 半自动化图像数据标注方法、电子装置及存储介质
JPH09293082A (ja) 画像検索装置及び画像検索方法
JPH09167246A (ja) 画像処理方法及び装置
CN114723601A (zh) 一种虚拟场景下的模型结构化建模及快速渲染方法
JP2012190434A (ja) 帳票定義装置、帳票定義方法、プログラム及び記録媒体
KR101118628B1 (ko) 지능형 인식 라이브러리 및 관리 도구를 활용한 고문서 이미지 데이터 인식 및 처리 방법
WO2025099621A1 (fr) Procédé semi-automatique de recherche d'images d'ouvrages de maçonnerie
CN119832570A (zh) 一种多模态分析的图像鉴别系统及方法
JPH0981583A (ja) ドキュメントファイリング装置および方法
JPH08194716A (ja) 画像処理方法及びその装置
CN112507999B (zh) 基于视觉特征的非侵入式用户界面输入项识别方法
CN117635763A (zh) 基于人像部件分析的自动换装方法、装置、设备及介质
KR20230103855A (ko) 이미지 스타일 변형 장치, 방법 및 이를 위한 컴퓨터 판독가능 프로그램
JP2004094427A5 (fr)
JP7075770B2 (ja) 文字認識システム、文字鮮明化システム、文字鮮明化プログラム、文字鮮明化方法、及び文字鮮明化画像の表示装置
Sokratis et al. A tool for tuning binarization techniques
CN119888234B (zh) 一种基于机器学习的古籍智能识别与修复方法及装置
JP3078038B2 (ja) 図面自動入力システム及びその方法
JP5402417B2 (ja) 画像処理装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24817713

Country of ref document: EP

Kind code of ref document: A1