Multi-scale model retrieval for case recommendation in manufacturability optimization

This paper proposed a multi-scale retrieval framework to help optimize manufacturability. The case based workflow integrates both geometry and rules with model-based definition. At conceptual stage, defect models are categorized and global retrieval is initialized to narrow down cases as defects references. At detail stage, local feature retrieval finds similar defect feature with pre-defined rules, to adapt to manufacturability optimization. Intersection feature detection and decomposition are explored to improve generality.