AIGC语境下纺织品纹样设计评价机制研究OA

The rapid advancement of artificial intelligence generated content(AIGC)has brought transformative changes to the field of textile pattern design.While AIGC tools enable efficient and diverse design generation,they also pose significant challenges in terms of identifying and evaluating AI-generated content.The boundary between human-made and machine-generated designs has become increasingly blurred,raising concerns about authenticity,cultural coherence,and industrial applicability.In response to these challenges,this study aims to establish a systematic evaluation mechanism for textile pattern design in the AIGC era.The primary objectives are to analyze the current capabilities and limitations of mainstream AIGC models in textile pattern generation,and to propose a comprehensive assessment framework that integrates technical,cultural,and production-oriented dimensions. This study begins with a critical review of the technical characteristics of representative AIGC generation models and their current applications in textile pattern design.Based on this analysis,three major challenges are identified:technical feasibility,cultural coherence,and production adaptability.To address these challenges,a multidimensional dynamic evaluation framework integrating"echnology-culture-production"is developed.The framework consists of a full-process assessment system that includes source file verification,detection of production defects,identification of generative artifacts,and cultural semantic crowd-evaluation.Quantitative indicators are also established in accordance with the Chinese industry standard CY/T 249-2021.To validate the proposed framework,an empirical study is conducted using traditional Chinese Yunjian patterns as a case study.Eight mainstream AIGC tools from both domestic and international sources are employed to generate patterns across eight distinct categories.The generated outputs are then systematically evaluated using the proposed framework and quantitative indicators.The empirical results demonstrate that the combined application of the full-process evaluation framework and the quantitative indicators effectively distinguishes AI-generated images from manually created ones.The study finds that while AIGC tools show considerable potential in generating visually appealing and structurally diverse patterns,they still exhibit significant limitations in maintaining cultural symbolism,production feasibility,and design coherence.In particular,none of the evaluated tools were able to fully replicate the nuanced cultural semantics and technical precision characteristic of human-crafted designs.The evaluation framework successfully identified key indicators of AI-generated content,including inconsistencies in pattern logic,traceable generative artifacts,and lack of cultural alignment. This study confirms that AIGC-generated content can be effectively identified and assessed through a systematic evaluation mechanism.The proposed multidimensional framework not only provides a reliable method for distinguishing AI-generated from human-made designs but also offers practical guidance for the quality control and industrial application of AIGC in textile design.The innovation of this study lies in its integration of technical,cultural,and production perspectives into a unified evaluation model,as well as its alignment with national industry standards.Furthermore,the study highlights the current limitations of AIGC tools in replacing the core functions of human designers,thereby reinforcing the irreplaceable value of human creativity and cultural expertise in the design process. As AIGC technologies continue to evolve,future research may expand the evaluation framework to include broader cultural datasets and more diverse design genres.Additionally,the development of automated evaluation tools based on the proposed indicators could further enhance the efficiency and scalability of content assessment.