The ability to perform automated feature extraction using AI detectors is a critical workflow for efficient digital twin generation. This functionality is already available in the standard/previous versions of the software.
The Issue: Currently, this capability is missing in the Flex version. This creates a significant "feature gap," forcing users who require advanced AI-driven workflows to revert to older versions or abandon the flexibility of the new platform.
Proposed Requirements:
Feature Extraction: Integrate AI-based detection tools to automatically identify and extract multiple object types from point clouds and meshes.
On-Prem Training Support: Provide the ability for users to train custom detectors locally (On-Prem).
Why this is essential:
Feature Parity: Users should not have to sacrifice high-level automation to move to the Flex platform.
Security & Privacy: On-prem training is essential for organizations with strict data governance policies that prevent uploading sensitive raw data to the cloud.
Customization: Local training allows us to develop detectors for niche elements specific to our regional standards or proprietary project requirements.
Business Value: Bringing this functionality to Flex will significantly accelerate adoption of the platform for large-scale infrastructure projects, where automation and data security are the primary drivers for technical decisions.
Regards,
Michael Kikinzon