Schematic structure of the research

Building your personal wardrobe

The processing flowof Impression Category Indicator

Experimental Results of garement features and category

Schematic structure of the Fashion Knowledge Visualization System

Fashion knowledge visualization system - 3D Style Cognition Space

Fashion knowledge visualization system - ||-coords-like Information Plot

Fashion knowledge visualization system - Elemenary Attribute Space

Digitalizing Fashion Knowledge – Extracting, Analyzing, Storing, and Application
〔數位化服裝領域知識之萃取、分析、儲存與應用〕

ABSTRACT

The research explores several issues in digitalizing fashion knowledge, such as extraction, storage, representation, and applications, from two different viewpoints: (1) general people who are not familiar with fashion but want an impressive and fashion outfit; and (2) novices at fashion who want to learn fashion efficiently. Two application systems are thus developed respectively, Intelligent Garment Impression Categorization System and Fashion Design Knowledge Visualization System. Hence, fashion knowledge, which is multidimensional and highly complex, is easy to be understood, transmitted, and reused using computer technologies.

The major functionality of Intelligent Garment Impression Categorization System is to build a personal digital wardrobe for each individual, and help users to find a garment fits a demanded impression category. For building a personal digital wardrobe, what a user needs to do is to take garment images and input to the system. The physical feature detector is then invoked to automatically extract essential data of garments, such as feature points and major color. Afterwards, several expressive features (e.g., cold degree, fancy degree, soft degree) are obtained using the expressive feature converter. Impression category indicator is exploited for assigning appropriate category terms to garments. Hence, an integrated two-tiered learning mechanism for categorizing garments into different impression terms is developed based on inductive learning.

Fashion Design Knowledge Visualization System aims to present abstract of apparel design, such as design styles and design elements, using technologies of information visualization. The system is schemed based on the framework of information visualization – information analysis, information presentation, and user interface interaction. Exemplars and knowledge of fashion design styles are firstly prepared and systemically analyzed by fashion experts. The interfaces are designed based on two viewpoints: fashion style and design elements. Style Cognition Space and Style Cluster View are designed for visualizing fashion style knowledge, while Comparison View of Design Elements Effects and Elementary Attribute Space are designed for visualizing knowledge of design elements. Several skills of graphic user interface and information visualization are used, such as Details on demand, Portals, and Zoom. Fashion knowledge thus can be analyzed and represented by directly or indirectly manipulating the developed visual objects.

Summary (中文總結)

本研究從兩個不同的使用者角度切入去探討服裝知識數位化的萃取、儲存、重現與應用，其一為對服裝知識沒概念但是卻想要擁有令人讚賞的穿著的一般使用者；另一個族群則是服裝設計領域的入門者，想要有效地學習服裝設計風格的相關知識。以此為基礎，兩個服裝設計領域知識數位化應用系統便在研究過程中因應而生－智慧型服裝意象分類系統與服裝設計知識視覺化系統。透過電腦科技的輔助，多維而且複雜的服裝設計知識得以輕易的被了解、傳遞以及重複使用。

智慧型服裝意象分類系統主要功能是在建置使用者個人數位化虛擬衣櫥並提供使用者可以從中搜尋一件符合特定感覺意象的衣服。為了因應普羅大眾使用，使用者只需輸入個人衣服數位影像，物理特徵偵測器便會自動偵測以擷取衣服必要的物理特徵，包括特徵點、主色調以及輪廓線等外在樣貌。接著，情感特徵轉換器會將所有物理特徵轉換成相對應的情感特徵。然後，感覺意象指定器則會將每件衣服單品依照專家定義的感覺意象加以分類。為了使本系統具備有專家定義的感覺意象分類知識，本系統應用了認知科學領域中歸納式學習(inductive learning)的精要，發展出一個二階層式服裝感覺意象語彙分類學習法，並透過與服裝專家們的合作，藉由問卷調查及深入訪談的方式取得專家對服裝感覺意象的共識成為演算所需的訓練樣本。

另一方面，服裝設計知識視覺化系統的發展是為了提供使用者一個輔助學習的工具，藉以了解服裝設計風格的專業定義，並且學習用不同設計元素組合加以呈現。本系統遵循了資訊視覺化領域裡的視覺化架構，資訊分析(information analysis)、使用者互動介面(user interaction interface)、和資訊呈現(information presentation)，發展出服裝設計知識視覺化系統，將錯綜複雜且多元化又抽象的服裝設計知識以具體的電腦影像呈現出來，作有效地視覺化呈現。服裝設計知識得以透過直接或間接地對視覺物件的選取和操控傳達給使用者並且有效地再被利用。

Challenges of Implementations

  1. Building Digital Personal Wardrobe

  2. Female Garment Impression Category Indicator

  3. Fashion Knowledge Visualization

Project Material

1. Fashion knowledge visualization system - 3D Style Cognition Space   (mp4, 853KB)

2. Fashion knowledge visualization system -  Style Cluster View  (mp4, 853KB)

3.Fashion knowledge visualization system -  ||-coords-like Information Plot  (mp4, 1.52MB)

4. Fashion knowledge visualization system -  Elementary Attribute Space  (mp4, 394KB)

Citation

1.   C.I. Cheng, D.S.M. Liu, and C.C.H. Lin, A digital tutor for learning fashion design, Multimedia Tools and Applications, 2014. (SCI, EI)

2.    C.I. Cheng and D.S.M. Liu, An intelligent clothes search system based on fashion styles, Proceedings of International Conference on Machine Learning and Cybernetics, Kunming, China, pp. 1592-1597, July 2008. (EI)