A method for constructing an electrofluid printing mode classification model
By constructing a classification model for current fluid printing patterns, utilizing a neural network framework and multi-layer encoder units, and combining a pre-training-transfer learning method, the problem of insufficient recognition accuracy in current fluid printing pattern monitoring in existing technologies is solved, achieving efficient pattern recognition and process parameter adjustment.
Patent Information
- Application Number
- CN202411174593.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2025-10-17
- Estimated Expiration
- 2044-08-26
AI Technical Summary
Existing methods for monitoring electrofluid printing patterns rely on manually extracting image features, which results in insufficient recognition accuracy and is easily affected by dynamic changes in the Taylor cone, leading to misjudgments.
A current fluid printing pattern classification model is constructed, including an image feature extraction module, an adjacent frame motion feature calculation module, a spatiotemporal flow feature extraction module, and a classification module. A neural network framework is used for pattern recognition, combined with a multilayer encoder and a multilayer perceptron output unit. The model is trained using a pre-training-transfer learning method and uses video data for pattern classification.
It improves the accuracy of pattern recognition, simplifies the calculation process, reduces computational consumption, and enables more efficient monitoring of current fluid printing, which can guide the efficient adjustment of process parameters.