Our Activity overview
This section outlines the activities of Astraea Software.
First, I will explain what the problems with current AI are, why we need 3D AI, and how we can use 3D AI.
Why do you need 3D AI?
We will talk about current AI issues, text-based, 2D data AI vs. 3D data AI, and why 3D AI is necessary.
At the sites of many industries such as the manufacturing industry and the construction industry, operations using 3D data are the mainstream. Computer-aided technology related to design and manufacturing called CAD, CAM, and CAE is also based on 3D shapes made by 3D CAD.
Therefore, when introducing AI to these technologies, it is essential for AI to recognize 3D shapes.
Comparison of 3D AI with text-based AI
Take database search, one of the main functions of AI, as an example.
With text-based AI, it is necessary to search for similar shapes by handing out textual information such as part numbers. However, it is necessary for a person to prepare the text information of the clue in advance assuming the search content, which requires careful preparation and enormous effort. Also, the maintenance requires a lot of effort.
If AI can recognize the 3D shape, it can search the database for a shape similar to the search target. Therefore, all you need to do in advance is to prepare only 3D shape data such as CAD.
Comparison of 2D AI and 3D AI
However, since there is no technology to recognize 3D shapes as they are, alternative measures have been taken, such as taking snapshots of the 3D shape from about 20 different viewpoints and having the 2D AI recognize the 3D shape from a group of 2D images.
This technology is an extension of 2D image recognition, and as the 3D shape becomes more complex, there are many cases where the hole shape is shadowed and not displayed on the screen, making it difficult to maintain recognition accuracy.
With the 3D AI technology that we have been researching and developing, the shape can be recognized directly from the nodes (vertices, coordinate positions) and elements (node connection information) that make up the 3D shape. Therefore, the shape of the hole etc. is not hidden, the dimensions can be grasped from the coordinate position, and it can be used for classification and matching of 3D shapes. Also, since it has coordinate positions, it is possible to generate new 3D shapes.
Furthermore, the 3D shape captured by the 3D scanner can be recognized as it is, and it can be classified and matched, and can be used in various scenes of the design process.
Fusion of AI / CAE / cloud
Our activity goal is to integrate CAE, AI, and cloud technologies and to infiltrate technologies from large companies to SMEs.
We will permeate the advanced technology cultivated in large companies into every corner of the industrial structure and strongly support the entire manufacturing industry all over the world.
We will create new added value by integrating 3D AI technology with CAE technology, which has already been used for decades at manufacturing design sites.
Furthermore, we provide a system with excellent operability, efficiency, and cost while reducing the burden on users with cloud technology.
The following image shows the configuration of the products we offer.
The 3D shape recognition AI models are provided under the product name Aries, and the Aries series consists of three products.
Includes a 3D AI model generator function. You can create an AI model by simply uploading your 3D shape.
Includes 3D AI model generator, 3D shape classify and 3D shape matching functions. The generator can generate two AI models, classify and matching, to classify mechanical parts such as bolts and brackets, and build a search system for similar parts.
Includes a 3D AI model generator and a 3D shape synthesis function.
The generator can be used to generate AI models for synthesis and build a conceptual design shape proposal system for products such as automobiles and consumer goods.
For the features of our main technology, 3D AI, please refer to the features of 3D AI models.