| Summary: | 碩士 === 國立成功大學 === 工程科學系碩博士班 === 97 === Electronic commerce in Internet is thriving. Even though consumers can effortlessly consume from the Internet, they still have troubles in finding products and commodities that will meet their needs and desires. The development of recommendation Systems is absolutely helpful for consumers to locate products that will meet their needs and desires in the messes of commodities. Recommendation Systems have been widely used for services of electronic commerce on any needs of real life, including books, restaurants, electronic games, tourist paths and so on, nevertheless, they are rarely used in fashion and outfit of clothes.
In real life, people may have difficulties in finding suitable outfits for different occasions, their own stature and favorite color combinations. They then try to solicit others or refer to fashion magazines for suggestions. In these ways, people gather information of specific clothing attributes, such as colors, brands, material, sizes, and types.
Accordingly, this thesis presents a contextual personalized wearing recommendation system. It creates a user’s personal wearing knowledge and help the user conveniently find suitable and favorite outfits for particular occasions through gathering users’ descriptions of specific clothing attributes, users’ preference and color images analysis. The results of experiments show the potential usability of this system in helping people in dressing.
|
|---|
