Dynamic Fashion Cultures
Now we can start a world-wide fashion analysis!
The paper presents a novel concept that analyzes and visualizes worldwide fashion styles. Our goal is to reveal web-based viral fashion styles. To achieve the fashion-based analysis, we created fashion culture database (FCDB), which consists of 76 million geo-tagged images in 16 cosmopolitan cities. By grasping a trend of mixed fashion styles, the paper also proposes an unsupervised fashion trend descriptor (FTD) using a descriptor, codeword vector and temporal subtraction. As the result of large-scale data collection and an unsupervised analyzer, we achieved world-level fashion visualization.
arXiv Pre-print, MIRU 2017 (Oral: 28.0%, Best Student Paper)
Kaori Abe (TDU/AIST), Teppei Suzuki (Keio/AIST), Shunya Ueta (Tsukuba/AIST), Akio Nakamura (TDU), Yutaka Satoh, Hirokatsu Kataoka (AIST)