| I&T Solution |
Artificial Intelligence Construction Drawings Vetting
(REF: S-0644) |
| Trial Project |
|
| Solution Feature |
- Our GUI enabled software enables the user to manage drawings by project and time, making it easier to track any changes throughout each update of the construction drawings.
- Using state-of-the-art RCNNs for object detection and transformer-based language generation technologies, our solution reads a series of construction drawings to identify changes, and express these changes with words.
- RCNNs, also known as Region-based Convolutional Neural Networks, are proven to have a higher accuracy as well as faster performance than conventional Convolutional Neural Networks.
- Should there be a need to expand to 3D construction drawings, RCNNs are extremely well suited to identify objects and text in "wild' environments such as photos
|
| Trial Application and Expected Outcome |
- We recommend to conduct the trial on the accuracy and efficiency of the RCNNs in identifying the differences in given construction drawings
- We believe the RCNNs are the key components of the project as they provide the analysis on the changes in the given construction drawings
- Once we prove the accuracy and efficiency of RCNNs, the main objective of the project will be accomplished - Analyzing construction drawings
- Although text generation is also a main objective of the project, the transformer model can simply learn from a given database of jargon, hence its accuracy can be maintained at a high standard
|
| Info on I&T Solution Provider |
| Solution Provider | : | Nysus Limited | | Address | : | Unit 206B2, 2/F. Tower 1, Harbour Centre, 1 Hok Cheung Street, Hunghom, Kowloon, Hong Kong | | Contact Person | : | Isaac To |
| Position | : | Head of AI | | Tel | : | 65808517 | | Email | : |
isaacto@nysus.io |
|
