CN-121433492-B - AR-based embroidery method, AR-based embroidery system, AR glasses and storage medium

CN121433492BCN 121433492 BCN121433492 BCN 121433492BCN-121433492-B

Abstract

The invention belongs to the technical field of embroidery and discloses an AR-based embroidery method, an AR-based embroidery system, AR glasses and a storage medium, wherein the AR-based embroidery method comprises the steps of inputting an embroidery sketch into a circulating countermeasure generation network to obtain an embroidery pattern, wherein the circulating countermeasure generation network is used for carrying out feature fusion on a line structure of the embroidery sketch and an embroidery pattern and generating the embroidery pattern; and in the embroidery process, responding to the acquired eye movement data and brain wave data, inputting the eye movement data and the brain wave data into an emotion model to obtain emotion quantification indexes, searching the emotion quantification indexes in a preset color-emotion association library, and obtaining and displaying recommended color combinations. Can realize convenient and efficient personalized embroidery creation assistance.

Inventors

LIU JIQIAN
Ai Honghua

Assignees

华南师范大学

Dates

Publication Date: 20260512
Application Date: 20251021

Claims (8)

1. An AR-based embroidery method, comprising: Inputting an embroidery sketch into a cyclic countermeasure generation network to obtain an embroidery pattern, wherein the cyclic countermeasure generation network is used for carrying out characteristic fusion on a line structure of the embroidery sketch and an embroidery pattern and generating the embroidery pattern; Embroidery is carried out by adopting AR glasses according to the embroidery pattern; In the embroidery process, responding to the acquired eye movement data and brain wave data, inputting the eye movement data and the brain wave data into an emotion model to obtain emotion quantification indexes, searching the emotion quantification indexes in a preset color-emotion association library, and obtaining and displaying recommended color combinations; The generator of the cyclic countermeasure generation network comprises a low-level network, a high-level network and a fusion network, wherein the low-level network is used for reserving a line structure of the embroidery sketch, the high-level network is used for extracting a gradient color mode, and the fusion network is used for realizing the combination of the line structure and the gradient color mode; The generator of the cyclic countermeasure generation network is further provided with a color attention mechanism for identifying color patch areas in the embroidery sketch and establishing an association between the color patch areas and the gradient color pattern.
2. The AR-based embroidery method of claim 1, wherein the emotion model includes a CNN branch for extracting an eye movement focusing time period and a pupil diameter variation from the eye movement data and an LSTM branch for extracting brain wave band energy from the brain wave data, and the emotion quantification index output by the emotion model includes pleasure and concentration.
3. The AR-based embroidery method of claim 1, wherein searching the emotion quantification index in a preset color-emotion association library to obtain a recommended color combination comprises: extracting the main tone of the current picture; Searching a preset color-emotion association library based on the main color and the emotion quantification index to obtain a plurality of color combinations; and screening out recommended color combinations from the plurality of color combinations by adopting a clustering algorithm.
4. The AR-based embroidering method of claim 1, further comprising, in an embroidering process: acquiring user data, wherein the user data comprises user behavior and/or preference data; Based on the user data, extracting a needle-method using sequence by adopting a latent semantic analysis and encoding the sequence into a vector; And inputting the vector into a mixed recommendation model to obtain recommended embroidery patterns and/or needle skills.
5. The AR based embroidery method of claim 4, wherein the hybrid recommendation model comprises a collaborative filtering module for filtering embroidery patterns with user scores below a set threshold, a content filtering module for recommending embroidery patterns matching the vector, and a feedback reinforcement module for dynamically adjusting recommended weights of embroidery patterns according to user selections.
6. An AR-based embroidery system, comprising: The embroidery pattern module is used for inputting the embroidery sketch into a cyclic countermeasure generation network to obtain an embroidery pattern, and the cyclic countermeasure generation network is used for carrying out characteristic fusion on a line structure of the embroidery sketch and an embroidery pattern and generating the embroidery pattern; the interaction module is used for embroidering according to the embroidery pattern by adopting AR glasses; The color combination recommending module is used for responding to the acquired eye movement data and brain wave data, inputting the eye movement data and the brain wave data into an emotion model, obtaining emotion quantification indexes, searching the emotion quantification indexes in a preset color-emotion association library, obtaining recommended color combinations and displaying; The generator of the cyclic countermeasure generation network comprises a low-level network, a high-level network and a fusion network, wherein the low-level network is used for reserving a line structure of the embroidery sketch, the high-level network is used for extracting a gradient color mode, and the fusion network is used for realizing the combination of the line structure and the gradient color mode; The generator of the cyclic countermeasure generation network is further provided with a color attention mechanism for identifying color patch areas in the embroidery sketch and establishing an association between the color patch areas and the gradient color pattern.
7. An AR glasses comprising a gesture recognition sensor for recognizing a gesture motion of a user, an eye tracking sensor for acquiring eye motion data of the user, an electroencephalogram sensor for acquiring brain wave data of the user, a memory storing executable program code, and a processor coupled to the memory, the processor invoking the executable program code stored in the memory for performing the AR-based embroidery method of any one of claims 1 to 5.
8. A computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute the AR-based embroidery method of any one of claims 1 to 5.

Description

AR-based embroidery method, AR-based embroidery system, AR glasses and storage medium Technical Field The invention belongs to the technical field of embroidery, and particularly relates to an AR-based embroidery method, an AR-based embroidery system, AR glasses and a storage medium. Background Embroidery takes needles as pens and threads as ink, colors and patterns are skillfully presented on the fabric through a plurality of exquisite needle skills (such as plain stitch, seed embroidery, plate gold embroidery and the like), and the embroidery has practical value and artistic ornamental value. In recent years, with the rapid development of Augmented Reality (AR) technology, its application in the field of traditional arts has been receiving attention, and an AR embroidery auxiliary system is one of typical applications thereof. At present, the existing AR embroidery auxiliary technology mainly focuses on providing an embroidery learning function for a user, namely, virtual information such as an embroidery step, a needle track and the like is superimposed on a real embroidery fabric or tool through the AR technology based on the existing fixed design pattern, and the user is guided to perform simulated embroidery according to the preset pattern and step. Although the mode reduces the threshold of embroidery to a certain extent, helps a beginner to quickly master basic embroidery skills, when a user hopes to design unique patterns according to own creative designs, adjust the layout and colors of pattern elements, or customize embroidery contents in combination with personal demands, the conventional system is difficult to provide effective technical support, the deep application of AR technology in an embroidery creation link is limited, and the demands of the user on personalized and creative embroidery works cannot be fully met. Disclosure of Invention The invention aims to provide an AR-based embroidery method, an AR-based embroidery system, AR glasses and a computer storage medium, which are used for realizing convenient and efficient personalized embroidery creation assistance. The first aspect of the present invention discloses an AR-based embroidery method, comprising: Inputting an embroidery sketch into a cyclic countermeasure generation network to obtain an embroidery pattern, wherein the cyclic countermeasure generation network is used for carrying out characteristic fusion on a line structure of the embroidery sketch and an embroidery pattern and generating the embroidery pattern; Embroidery is carried out by adopting AR glasses according to the embroidery pattern; In the embroidery process, responding to the acquired eye movement data and brain wave data, inputting the eye movement data and the brain wave data into an emotion model to obtain emotion quantification indexes, searching the emotion quantification indexes in a preset color-emotion association library, and obtaining and displaying recommended color combinations. In some embodiments, the emotion model includes a CNN branch for extracting an eye movement focusing time length and a pupil diameter change from the eye movement data, and an LSTM branch for extracting brain wave band energy from the brain wave data, and the emotion quantification index output by the emotion model includes pleasure and concentration. In some embodiments, the searching the emotion quantification index in a preset color-emotion association library to obtain the recommended color combination includes: extracting the main tone of the current picture; Searching a preset color-emotion association library based on the main color and the emotion quantification index to obtain a plurality of color combinations; and screening out recommended color combinations from the plurality of color combinations by adopting a clustering algorithm. In some embodiments, the generator of the cyclic countermeasure generation network includes a lower level network for retaining the line structure of the embroidery sketch, a higher level network for extracting the gradient color pattern, and a convergence network for realizing a combination of the line structure and the gradient color pattern. In some embodiments, the generator of the cyclic countermeasure generation network is further provided with a color attention mechanism for identifying color patch areas in the embroidery sketch and establishing an association between the color patch areas and the gradient color pattern. In some embodiments, during the embroidery process, further comprising: acquiring user data, wherein the user data comprises user behavior and/or preference data; extracting a needle usage sequence according to the latent semantic analysis based on the user data and encoding the needle usage sequence into a vector; And inputting the vector into a mixed recommendation model to obtain recommended embroidery patterns and/or needle skills. In some embodiments, the hybrid recommendation model comprises a collaborative