Construction method of electrofluid printing mode classification model

By building a neural network framework, using the time and spatial flow characteristics in video data, the precise classification of the current fluid printing mode is achieved, and the problems of insufficient identification accuracy and dynamic change monitoring in the existing technology are solved, and the current fluid printing mode is improved recognition ability.

CN119169346AActive Publication Date: 2024-12-20HUAZHONG UNIV OF SCI & TECH
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411174593.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-12-20
Estimated Expiration
2044-08-26

AI Technical Summary

Technical Problem

The existing classification methods for galvanomic printing mode rely on manual extraction of image features, lack of recognition accuracy and difficult to effectively monitor dynamic changes of Taylor cone mode.

Method used

Building a neural network framework, including image feature extraction module, adjacent frame motion feature calculation module and spatiotemporal stream feature extraction module, and video data extracting time and spatial stream features, combining multi-layer encoder and multi-layer perceptron output unit to realize Accurate classification of galvanomic printing modes.

Benefits of technology

The classification accuracy of the current-fluid printing mode is improved, the calculation process is simplified, the calculation consumption is reduced, the pattern recognition ability of dynamic changes is enhanced, and the problem of insufficient recognition accuracy of the image-based classification model is compensated.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119169346A_ABST
    Figure CN119169346A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of electrofluid printing, and particularly relates to a method for constructing an electrofluid printing mode classification model, which comprises the following steps: each training sample comprises an electrofluid Taylor cone and jet segment video containing f frames of images and a corresponding electrofluid printing mode; an image feature extraction module is used for preliminarily extracting image features; the adjacent frame motion feature calculation module is used for calculating feature-level motion information between adjacent frames, the time stream feature extraction unit is connected with the image feature extraction module and is used for further extracting motion features, and the spatial stream feature extraction unit is connected with the adjacent frame motion feature calculation module and is used for further extracting image features; the classification module is connected with the time stream feature extraction unit and the spatial stream feature extraction unit for classification. According to the method, the video data is used as input and is combined with a space-time double-flow attention mechanism, so that the problem of insufficient recognition precision of an image-based classification model is effectively solved.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Information processing method, image display method, information processing device, and image display device

    CN112313701A

  • Taylor cone dynamic identification and detection method based on near-field electrospinning direct writing of machine vision

    CN112634198A

  • Langmuir circulation observation control method and system

    CN114842250A

  • Point of view video processing and curation platform

    US20200349977A1

  • Inspection system for three-dimensional printer and methods thereof

    US20240173972A1