What Google does in indexing 2D data (such as text and images) is exactly what Physna’s Thang search engine does in processing physical objects and 3D models. Physna’s software, established in 2016, is bridging the gap between the real world and digital code, and organizes the 3D world through a deeper understanding of the physical properties of real world objects and the relationships between them. Physna is changing the future volume of engineering, industrial design and procurement.
Sequoia Capital made a Series B investment in Physna in January 2021 (the company’s total financing reached 29 million US dollars at the time). After half a year, Physna received a new round of 56 million US dollars in financing, bringing Physna’s total financing to 8600 Ten thousand U.S. dollars. The new capital was led by the venture capital firm Tiger Global, with participation from Alphabet Inc.’s venture capital arm GV (Google Venture Capital) and Sequoia Capital.
© Physna
Google for 3D physical world search
Generally speaking, it is difficult for us to use CAD or3D scanningThe constructed 3D model of physical entity retrieves the physical object, especially when the retrieval target is a part of the larger component, the algorithm is more difficult to maintain.
In other words, if you want to find a specific target from the 3D data, it is very complicated to directly use the model matching method. It is like we scanned the 3D data of a car, but now we want to use a 3D model of a screw to match the car wheel. This process is very complicated and time-consuming.
Traditional solutions are expensive and time-consuming, which not only affects the engineering procurement process, but also affects component identification. It is estimated that more than 70% of the economy is centered on physical objects, and less than 1% of software can handle 3D data.
Physna’s CEO Paul Bowers and CTO Glenn Warner founded Physna in 2015 with the original purpose of protecting the intellectual property rights of product designs from theft. But in 2016, the company turned, hoping to use deep learning technology to encode 3D models into software understandable data, bridging the gap between the real world and the digital world.
Ternary code
The initial inspiration came from the use of Morse code to process information in dots and dashes in telecommunications, followed by binary codes. These two symbol systems laid the foundation for calculations. Physna’s software is similar to a “ternary code”. In the field of ternary codes, computers can understand objects in natural forms without human intervention.
Physna will decompose the structure of the 3D model and analyze each part to determine the relationship between the different models. Customers can find models by searching for 3D models, partial models, geometric dimensions or even just model data. *
© Physna
Through geometric deep learning and mathematical analysis, Physna can compile 3D models into data that can be read by software. Users can search using custom 3D objects, models, and geometric data to find brand new 3D models.
Physna’s AI uses predictive descriptions, classifications, costs, and materials to find matching models. The platform displays all repetitive or similar parts, even components in complex objects, as well as the exact position and quantity of the components in the model.
More precisely, Physna can help you quickly find and match similar parts, even parts in assembly, replacement parts for this part, and alternative use of the part, and can identify parts based on their shape.
If there is no 3D model or part number, try to describe some features (such as the estimated value of the hole diameter); if there is no CAD drawing, then try to provide some detailed information about the model and specifications (including materials, cost information, classification, and processing Sex and suppliers). In these cases, it can almost help you retrieve the parts you want.
Finally, after engineers mark metadata and information, Physna can continue to learn from existing models. In addition to production forecasts, in addition to estimating costs, materials, and manufacturability, the technology can predict part performance on the surface based on historical data and point out potential design flaws.
Search Engine Thans
Physna’s search engine Thangs brings industry-leading search and collaboration capabilities to the world of 3D data, allowing hardware developers to access important resources that software developers can use for a long time. Google’s work on indexing 2D data (such as text and images), and Thanngs’ work on handling physical objects and 3D models. Thangs allows users to find parts based on their (potential) relationship with other parts. For example, just by uploading a 3D model, Thanns users can find out where the part can be used, or which parts can be obtained from a supplier suitable for the model as a component.
© Physna
Thangs also supports text search. If you try to search for “car” in the search box of Thangs, you will be able to search for 3D models of many cars. These model data come from many third-party platforms. After clicking, you will be redirected to the corresponding website.
Thangs can quickly find 3D models, identify them accurately, understand the relationship between models and components, and make their 3D models smarter. In addition to the powerful search function, Thangs also introduced collaboration tools that 3D model creators have been looking forward to, such as automatic version control and easy collaboration on team projects.
Thangs can find compatible parts and suppliers. Thangs analyzes the physical properties of each 3D model and displays its relationship with other 3D models online, including parts suitable for its assembly and partial design. It even makes predictions that reveal relevant parts suppliers.
Physna’s current users include the US Department of Defense and a large number of Fortune Global 100 companies, covering multiple industries such as aerospace, automotive, manufacturing, medical and robotics.
Physna enables the software to truly understand physical three-dimensional data, thus achieving a huge technological leap. From geometric search to 3D machine learning and prediction, through the integration of physics and digital, Physna unlocks more opportunities and development space.
(Editor in charge: admin)
0 Comments for “Searching for 3D models, Physna gets a new round of US$56 million investment from Sequoia and Google”