Features

Home
Features
Artificial Intelligence

Artificial Intelligence

As developing the solution required various features and intelligence, We have applied the technology of data-based and rule-based machine learning, which is a kind of AI, and developed the optimal solution through it.

Automatically generating the apartment size and shape image

We developed an automatic image generating module for creating thumbnails of our own 30,000 apartment size and shape data.

  1. The shape information of construction plan(ex. Area, Shape of Room, Opening, etc.) The automatic identification of the usage of space through rules (ex. Room, Toilet, Living Room, Terrace, Porch, Warehouse, Duct, etc.)
  2. The automatic generation and alignment of the entire outer dimension line
  • 이미지
    Automatically generating
         1 / 4     
  • 이미지
    Automatically generating
         2 / 4     
  • 이미지
    Automatically generating
         3 / 4     
  • 이미지
    Automatically generating
         4 / 4     
  • 이미지
    Automatic Store Arrangement
         1 / 3     
  • 이미지
    Automatic Store Arrangement
         2 / 3     
  • 이미지
    Automatic Store Arrangement
         3 / 3     

Automatic Store Arrangement

We developed an automatic placement simulator for each type of store such as Samsung Electronics TV, Samsung Electronics Mobile, LG Electronics TV, etc.

  1. The amount, location, and direction of products and racks are determined by rules made by size and shape of Highlight zone, entrance, copper wire, wall, etc.
  2. The optimized solution is provided for the best exhibition presentation within the limited space of the store

Algorithm for creating space mesh

It create 3D meshes automatically based on the CSG(Constructive Solid Geometry) Algorithm.

  1. We have our own algorithm of Triangulation for creating surfaces that are made up of diverse lung curves and openings
  2. Auto-processing of the diverse shape of finishing materials created by the composition of space and connection between different bodies of walls under the rule of construction space
  • 이미지
    Algorithm for creating space mesh
         1 / 2     
  • 이미지
    Algorithm for creating space mesh
         2 / 2     
이미지
The shortest distance algorithm
     1 / 1     

The shortest distance algorithm

We use the Shortest Distance Algorithm, Dijkstra Algorithm, for developing modules of direction guidance in the interior space.

  1. Suggesting a guide of optimal direction through distance information and the nodes of each important position(ex. holes, corridors, corners, entrances, etc.)
  2. Developing a module of guiding direction based on data through real-time change management of nodes in which space changes

Object tracking

It processes learning about object tracking in the apartment size and shape through machine learning.

  1. It is possible to track the object concerning facilities such as rooms, toilets, and entrances, and facilities such as doors, windows, and elevators
  2. It is utilized as fundamental data such as the classification of space usage, opening location, automatic 3D generation, etc.
이미지
Object tracking
     1 / 1     
이미지
The automatic generation of 3D
     1 / 1     

The automatic generation of 3D the aprtment size and shape data

We developed automatic generation module for creating 3D apartment size and shape image according to 2D apartment size and shape image.

  1. Machine-Learning-based information of analyzing 2D image through wall, room, door, and window Creating the size and shape files of KOVI-Archi classified by objects. (more than 90% accuracy)
  2. Automatically adjusting scales by comparing them with using space by room objects. (more than 95% accuracy)
  3. Automatically arranging basic furniture and materials by the usage of room

Tracking indoor locations based on reinforcement learning

We developed the module for tracking indoor location through reinforcement learning, one of the machine learning techniques.

  1. It processes reinforcement learning about tracking location based on the reward system that scores the distance between the actual locations according to moving direction(Action), signal value(Input) of RSSI(Received Signal Strength Indicator) of AP(Access Point)
  • 이미지
    Tracking indoor locations
         1 / 3     
  • 이미지
    Tracking indoor locations
         2 / 3     
  • 이미지
    Tracking indoor locations
         3 / 3     

Related paper

  • Reinforcement Learning-Based Interior Location Tracking 이미지

TOP