CADModelScope: Revealing the Dependency Structure Behind Parametric Computer-Aided Design Models
Authors
Paper Title
CADModelScope: Revealing the Dependency Structure Behind Parametric Computer-Aided Design Models
Publication Info
- Topic area: Visualization and navigation of parametric CAD model dependencies.
- Keywords: Parametric CAD, dependency visualization, modularization, design intent, graph-based tools, CAD navigation, operation dependencies, user study, Autodesk Fusion, CAD debugging.
Background and Problem
- Problem / challenge: Parametric CAD models rely on hidden interdependencies among operations, which are obscured in commercial CAD systems. This makes navigation, debugging, and modularization of complex models difficult and error-prone.
- Significance: Understanding and managing operation dependencies is critical for efficient design workflows, error resolution, and collaboration in CAD, especially for large and unfamiliar models.
- Motivation and related work: Prior efforts to visualize dependencies in CAD systems have been limited by visual clutter, lack of interactivity, and insufficient user-centric evaluations. Existing tools do not effectively support tracing, modularization, or error resolution in complex models.
Solution
- Proposed approach: CADModelScope, a multi-level graph-based visualization tool integrated into Autodesk Fusion, reveals hidden operation dependencies to support navigation, modularization, and debugging.
- Novelty:
- Multi-level visualization of operation dependencies, including global, local, and modular views.
- Clustering of interdependent operations into functional modules.
- Interactive synchronization with the native CAD interface for seamless navigation.
- Empirical evaluation of the tool's impact on user workflows.
- Procedure and key techniques:
- Construct a directed acyclic graph of operation dependencies, with nodes representing operations and edges indicating dependencies.
- Implement three views:
- Operation Overview: Global dependency graph with node size indicating downstream influence.
- Local Dependency View: Interactive tracing of upstream and downstream dependencies for a selected operation.
- Modular View: Clustering of operations into functional units using the Louvain algorithm.
- Synchronize the tool with Autodesk Fusion’s UI for real-time interaction and highlighting.
- Evaluate the tool through a user study with tasks reflecting navigation, modularization, and error resolution.
Results
- Concrete findings:
- CADModelScope improved structured navigation and modularization of CAD models.
- Participants using the tool achieved higher accuracy in modularization tasks (mean accuracy = 0.859 vs. 0.761 without the tool).
- Dependency-aware workflows enabled targeted error resolution and reduced reliance on exhaustive search.
- Advantage over baselines:
- Reduced cognitive load and navigation overhead compared to native CAD features.
- Enabled context-aware exploration of dependencies, which was not possible with traditional tools.
- Experiments / evaluation:
- Conducted with 8 participants (mechanical engineers with CAD experience).
- Tasks included navigation, modularization, and error resolution in complex CAD models.
- Mixed-method evaluation combining task performance metrics and qualitative feedback.
- Limitations and future work:
- Limited to part modelling; future work could extend to assemblies.
- Clustering results occasionally misled users; future iterations could integrate user-defined semantics.
- Study used publicly available models and a modest participant sample; larger-scale and longitudinal studies are needed for generalizability.
Summary
CADModelScope addresses the challenge of visualizing hidden dependencies in parametric CAD models by providing a multi-level graph-based tool integrated into Autodesk Fusion. The tool enables structured navigation, modularization, and error resolution through interactive dependency visualizations. A user study demonstrated its effectiveness in improving workflows compared to native CAD features, particularly for complex and unfamiliar models. Future work could extend the tool to assemblies, incorporate user-defined semantics, and evaluate its impact in more diverse and realistic design contexts.
Research Questions / Practical Problems
Question signals indexed for this paper.
- 67%
Super-Resolution Capacitive Touchscreens
CHI '21· Vibrotactile Feedback & Skin Stimulation +1
- 60%
Interactive Document Clustering Revisited: A Visual Analytics Approach
IUI '18· Interactive Data Visualization
- 60%
Data-centric disambiguation for data transformation with programming-by-example
IUI '21· Interactive Data Visualization
Based on Jaccard similarity of research subtopics & professions (≥60%)