KABEPITA, an AI application that identifies wallpaper part numbers in seconds, developed through industry-academia collaboration App Store (January 31), Google Play (February 7) App Store (Jan. 31) and Google Play (Feb. 7)

Equipped with "Automatic Texture Identification Program" Jointly Developed with Doshisha University Contributes to construction DX by reducing hours of work to a fraction of a second

2024.02.07 11:00

コマツ株式会社

Komatsu Ltd. (Head Office: Higashi-Osaka City, Osaka; President: Satoshi Komatsu) has launched the wallpaper AI identification application "Kabe Pita," which is equipped with an "automatic texture identification program" jointly developed with the Intelligent Mechanism Laboratory (Professor: Masahiro Okuda), Department of Intelligent Information Engineering, Faculty of Science and Engineering, Doshisha University, on the App Store (January 31) and Google Play (February 7), The application is available free of charge for the time being and can be used for triangulation.

The application is available free of charge for the time being, and can be used three times a day on a trial basis, or an unlimited number of times by registering as a free member. We will continue to improve and debug the functionality of the system as it is used by a wide range of users, including those in the construction industry.

Wallpaper icon

The wallpaper product number is written on the reverse side of the product, which is the adhesive side, due to the characteristics of the product, making it difficult to identify the product number, especially for popular wallpaper (mass-produced cloth), which has many similar products, making it difficult even for skilled craftsmen to distinguish between them. In the interior decoration industry, it was common practice to search for the same or similar product numbers at the request of customers, which required long hours of labor. However, by using "KABEPITA," the identification process can be completed in a few seconds.

Main Functions of KABEPITA

KABEPITA] List of Functions

KABEPITA instantly identifies approximately 600 types of popular wallpaper (mass-produced cloth) from six major Japanese manufacturers (Sangetsu, Shinkor, Tori, Tokiwa, Lilycolor, and Lunon) by simply photographing them with a smartphone, and displays the top five manufacturers, part numbers, and catalogs.

The probability of identification is close to 90% for the topmost product and 95% for the top five product numbers. For discontinued product numbers and product numbers that have not been databased (wallpaper other than popular wallpaper from the six major domestic manufacturers), the five closest product numbers from among the listed product numbers are displayed.

This application not only improves the efficiency of identification work, but is also very useful for limiting the scope of installation, selecting repair materials, and assisting in design, planning, and coordination.

Detailed information on the "KABEPITA" application

<iPhone version

Title: "KABEPITA" iPhone application

Recommended environment: iOS version 6.0 or higher

Price: Free of charge

Location: App Store

URL ： https://apps.apple.com/jp/app/%E3%81%8B%E3%81%B9%E3%81%B4%E3%81%9F/id6476473697

App Store 2D code for download

＜Android version

Title : Android application "KABEPITA

Recommended environment : Android OS version 5.0 or higher

Price ： Free

Location : Google Play

URL ： https://play.google.com/store/apps/details?id=com.textoragekabe.pita&pli=1

Google Play 2D code for download

KABEPITA application official page

https://textorage.com/kabe-pita/

KABEPITA Pamphlet

What is an "automatic texture identification program"?

Currently, most of the mainstream photographic identification software analyzes various factors such as color, shape, background, etc., in order to identify an object comprehensively. On the other hand, "automatic texture identification programs" are capable of identifying the manufacturer and part number of a product based solely on the local (or partial) characteristics of the texture, or material part.

By using "automatic texture identification programs" in combination with existing identification programs, various issues that were previously impossible to identify may be eliminated.

The technology can be applied not only to interior materials, but also to fabrics, leather, stone, and many other materials.

This technology was presented at the international conference [2022 IEEE 11th Global Conference on Consumer Electronics], and a joint patent application has been filed by Komatsu Ltd. and Doshisha University. ([Patent Application No.] Patent Application 2023-150356)

Explanation of automatic texture identification program at the presentation of new KABEPITA products

Prof. Masahiro Okuda Intellectual Mechanisms Laboratory

https://vig.doshisha.ac.jp/