請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/76995| 標題: | 消費者於線上購買服飾時採用虛擬試衣技術之探討:以人工智慧Deepfake技術應用為例 Analyzing Consumers’Adoption of Virtual Try-on Technology for Online Apparel Shopping: Application of AI Deepfake Technology as an Example |
| 作者: | CHIA YI LIN 林珈伊 |
| 指導教授: | 黃恆獎(Heng-Chiang Huang) |
| 關鍵字: | 虛擬試衣間,人工智慧,深偽,科技接受模型,創新擴散理論,服務創新,遊戲化,服務流程遊戲化, Virtual Try-on,AI (Artificial Intelligence),Deepfake,TAM(Technology Acceptance Model),IDT(Innovation Diffusion Theory),Gamification, |
| 出版年 : | 2020 |
| 學位: | 碩士 |
| 摘要: | 科技日新月異下,許多商業模式也跟著轉變,尤其AI的出現改變了整個IT產業外,也改變了整個世界。在這個快速變遷的世代,迫使許多企業開始思考如何使用這些新科技來適應這個變遷。AI觸發軟體開發流程的改變,讓許多商機應運而生,而其中對於服務產業來說AI所帶來的一大商機「資料提供與處理」透過大量標籤資料來支援AI做學習及決策能提供消費者突破沒有資料的瓶頸,而這其中也影響了許多產業順著這個趨勢做出演變,例如對於服務流程的影響。 例如過去服飾產業雖然也衍生出線上採買的流程,然而最大的矛盾點還是在於消費者無法想像無法接觸到的產品穿在自身上的模樣,而AI的Deepfake技術利用了卷積神經網絡(CNN或ConvNet)—一種深度學習神經網絡來能彌補這一個缺口。透過Deepfake能夠實現所謂的“少量學習”功能,即能夠在僅合成少量圖像的情況下進行“學習”並進行自我訓練,然後再合成全新的圖像,人工生成的圖像。實際上,該系統還具有“一次性學習”的能力,儘管僅添加一個圖像,但它可以從一個原始圖像中生成合理的結果,添加更多圖像可以提高最終表示的準確性。因此本研究將從AI的Deepfake角度來探討虛擬試衣間對於消費者以及產業之影響。 With the rapid development of science and technology, many business models have also changed. In particular, the emergence of AI has changed the entire IT industry as well as the entire world. In this rapidly changing era, many companies are forced to start thinking about how to use these new technologies to adapt to this change. AI triggers changes in the software development process, giving rise to many business opportunities. Among them, for the service industry, AI brings a major business opportunity 'data provision and processing' through a large amount of label data to support AI learning and decision-making can provide consumption They break through the bottleneck of no data, and this has also affected many industries to follow this trend to evolve, such as the impact on service processes. For example, although the clothing industry in the past has also derived the online purchasing process, the biggest contradiction is that consumers cannot imagine the appearance of untouchable products on themselves. AI's Deepfake technology uses convolutional neural networks (CNN or ConvNet)-a deep learning neural network to make up for this gap. Through Deepfake, the so-called 'small amount of learning' function can be realized, that is, the ability to 'learn' and conduct self-training while only synthesizing a small amount of images, and then synthesize brand new images, artificially generated images. In fact, the system also has the ability of 'one-time learning'. Although only one image is added, it can generate reasonable results from an original image. Adding more images can improve the accuracy of the final representation. Therefore, this study will explore the impact of virtual fitting rooms on consumers and the industry from the perspective of AI's deepfake. |
| URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/76995 |
| DOI: | 10.6342/NTU202001987 |
| 全文授權: | 未授權 |
| 顯示於系所單位: | 事業經營碩士在職學位學程 |
文件中的檔案:
| 檔案 | 大小 | 格式 | |
|---|---|---|---|
| U0001-2807202017430500.pdf 未授權公開取用 | 4.23 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。
