Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | School of Fashion and Textiles | en_US |
| dc.contributor.advisor | Li, Li (SFT) | en_US |
| dc.creator | Wu, Xiaopei | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/13944 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | The application of deep learning driven computational creativity in knitted textile design | en_US |
| dcterms.abstract | In recent years, artificial intelligence (AI) has proliferated into an expanding array of interdisciplinary applications across various industries and academic fields. In particular, it has been increasingly applied as a tool to facilitate or exhibit creativity. When it comes to the fashion & textiles industry, while there has been an evident movement towards digital tools and automation, and a recognition that AI applications can benefit productivity, however it is ultimately found that the practical, scaled adoption of AI is skewed towards the business, manufacturing and retail functions of the industry, and away from the creative design functions, where AI applications remain largely experimental and hypothetical. Furthermore, existing studies of AI applied to creative design functions of the fashion & textiles industry are largely focused on outward design elements such as style, silhouette, color and pattern, and lack in addressing inward textile material elements such as structure, composition, and finish. Furthermore, the outcomes of existing studies manifest a lack of practical applicability to the real-world situation of the creative design process in the fashion & textiles industry. | en_US |
| dcterms.abstract | Therefore, to address this gap, this study aims to explore the practical and creative potential of AI, in the form of deep learning driven computational creativity, to facilitate the textile design component of the real-world creative fashion design process. To achieve this, the study addresses three research objectives (RO) as follows: (1) to systematically explore the research and methods surrounding the topic of deep learning driven computational creativity for knitted textile design in the fashion creative design process (which is encompassed in the literature review section), (2) to devise a deep learning based computational method that facilitates the knitted textile design component of the knitted fashion creative design process (which is exhibited by the methodology and results section), (3) to evaluate empirically the creative value and practical utility of the devised computational method (which is exhibited by the conclusion section). Consequently, the scope of this study encompasses the intersection of 3 key topic areas: deep learning, knitted textiles, and the creative design process. Each of these 3 key topic areas are interpreted as respective subfields of: Computing (AI), Fashion (Textiles), and Creativity (Creative Practices). The intersection of these 3 themes represents the gap manifested in existing literature which this study aims to tackle. | en_US |
| dcterms.abstract | In terms of methodology, RO(1) was conducted by a systematic thematic literature review, with both academic and industry sources, RO(2) adopted a practice-led approach to devise the applied computational method intended to facilitate knitted textile design in the context of the creative fashion design process, by involving digital programming and design tools and physical knitting machinery tools, and RO(3) used a mixed method of surveys & interviews to evaluate the devised computational method applied, with general public and fashion industry experts respectively. | en_US |
| dcterms.abstract | The outcome produced by RO(1) was a contextual perspective of the intersection between this study’s 3 key themes (deep learning, knitted textiles, and the creative design process), which subsequently indicated a research gap in textile-focused creative fashion design applications of AI that have real-world, practical utility, and that are devised to facilitate the design process, not just execute product creation. The result of RO(2) was an AI driven computational method (based on a deep learning StyleGAN model), able to generate recognizable knitted textile images with distinguishable physical attributes (such as stitch, gauge and hand-feel), that proved to be sufficient visual reference for factory knitting technicians to interpret into corresponding physical swatch samples. The results from the evaluative surveys and interviews of RO(3) suggest that the devised AI driven computational method could contribute creative and practical value if adopted as an assistive tool to facilitate knitted textile design in the creative fashion design process. Furthermore, the survey results indicate that the AI generated knitted textile images were viewed by participants with the same level of discernment towards the physical and aesthetic attributes as the real ones. Then when projected onto garments, the generated knitted designs were rated overall more creative, fashionable, and buyable than the real ones. The findings from the interviews with fashion industry experts indicate a collective openness to AI facilitating the creative fashion design process, but only as an assistive tool to shoulder the non-creative tasks of the design process, such as deciphering data on customer needs and market trends or for providing design inspiration that empowers designers to do their creative tasks more effectively. The interview results also indicate an overriding consensus that the absence of tactility is a major barrier to embracing any form of digital based tools for knitted textile design, according to the design creatives. | en_US |
| dcterms.abstract | The outcomes of this study have both theoretical and practical significance. The theoretical significance manifests in the novelty of applying a GAN based deep learning model towards the textile component (specifically, towards the uniquely spatial attributes of knitted textiles) of the fashion design process, and in the novelty of the empirical methodological approach which took into consideration the realities and challenges of the creative fashion design process. In terms of practical significance, this study has shown that if the devised AI driven computational method was applied in a real-world fashion industry context, it has potential to: (a) augment creativity of designers by providing data-driven textile design inspiration and alleviating non-creative tasks, (b) increase operational efficiency by reducing workload, (c) be digitally compatible for an increasingly digital fashion design process, (d) foster sustainability by reducing waste and sampling needs, (e) reduce development lead-times and costs, (f) inform customer needs and trends by learning from the dataset. | en_US |
| dcterms.abstract | All in all, this study’s application of deep learning driven computational creativity in knitted textile design has the potential to positively impact and evolve the fashion & textiles industry’s creative designers, process, and product. | en_US |
| dcterms.extent | 340 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2023 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/13944