MCFR: A Mask-Guided Coarse-to-Fine Regression Framework for Robust Multi-Variant Board-to-Board Connector Assembly
It demonstrates an average real-world insertion success rate of 99.25% for multi-variant board-to-board connectors in smartphone assembly by combining mask priors with photometric refinement.
The authors introduce MCFR, a mask-guided coarse-to-fine regression framework designed to automate board-to-board connector insertion across multiple component variants in 3C manufacturing. The approach uses an object-aware mask prior and explicit photometric refinement to suppress background interference and improve alignment robustness. Evaluated on a multi-variant dataset, a batch insertion testbed, and a real smartphone assembly task, MCFR outperformed representative baselines.
Client angle, caveats & evidence
Client angle Relevant if a client is seeking to automate high-precision visual alignment and connector insertion across varying component designs in 3C manufacturing.
Caveats The framework is validated on a self-constructed dataset, a dedicated testbed, and a single smartphone assembly task rather than broad standardized public benchmarks.