Generative AI-assisted CAD Modeling of Electronic Enclosures

Conversational AI for Parametric Design

This project focuses on developing a cutting-edge methodology to convert high-resolution Signed Distance Fields (SDF) or triangle meshes into Boundary Representation (B-Rep) models tailored explicitly for electronic enclosure components. By leveraging advanced annotations from the ABC dataset, the foundational Scan-to-BRep model will be refined to enhance boundary and corner accuracy, creating a robust solution for handling intricate geometries of Electronic enclosures.

🏛️ BITS Pilani, Hyderabad Campus

Platform Overview

Multi-Modal CAD Interface is a web-based conversational AI assistant designed to streamline human-AI interaction for parametric CAD modeling. The platform supports multi-modal input approaches combining text, images, and camera inputs to create CAD models through natural dialogue.

Key Features

  • Multi-Modal Input Support: The system supports text prompts, reference images, and OpenMV camera integration for real-time object capture and CAD model generation.
  • Interactive 3D Viewer: Features a Three.js-powered viewer with topology visualization, allowing users to inspect faces, edges, and vertices of generated CAD models.
  • BREP Manipulation: Real-time geometric operations including face extrusion and boolean operations using OpenCASCADE for precise parametric design workflows.
  • LangChain Orchestration: Advanced workflow management with chain-of-thought reasoning and seamless integration between frontend inputs and backend CAD generation agents.

Platform Demonstrations

Explore how the Multi-Modal CAD interface works in practice through these demonstration videos. See the platform in action with various input methods including text prompts, image references, camera integration, and real-time editing. Each demonstration showcases different aspects of the conversational CAD workflow.

Text-to-3D Generation

Demonstration of generating 3D models from natural language descriptions, showing the complete workflow from text input to 3D model generation.

OpenMV Image-to-3D Generation (Demo 1)

OpenMV camera demonstration showing real-time image capture and conversion to 3D models, showcasing the complete workflow from camera input to CAD generation.

Project Assets & Resources

Explore the comprehensive collection of assets, tools, and resources utilized throughout the Multi-Modal CAD interface development. From UI mockups and wireframes to technical diagrams, 3D model examples, and development screenshots. Each asset category showcases different aspects of the project's visual and technical documentation.

OpenMV Camera Captures

OpenMV capture 1
OpenMV capture 2
OpenMV capture 3
OpenMV capture 4

Frontend UI Screens

UI Screen 1
UI Screen 2
UI Screen 3

Live Demo Screens

Demo Screen 1
Demo Screen 2
Demo Screen 3

Project Contributors

Vedant Agarwal
Tushar Anand
Pritham Kumar Jena
Bhavika Baburaj
×