PROJECT TITLE :

Mining Fashion Outfit Composition Using an End-to-End Deep Learning Approach on Set Data - 2017

ABSTRACT:

Composing fashion outfits involves deep under-standing of fashion standards while incorporating creativity for choosing multiple fashion items (e.g., jewelry, bag, pants, dress). In fashion websites, in style or high-quality fashion outfits are typically designed by fashion experts and followed by giant audiences. In this paper, we have a tendency to propose a Machine Learning system to compose fashion outfits automatically. The core of the proposed automatic composition system is to get fashion outfit candidates primarily based on the appearances and metadata. We have a tendency to propose to leverage outfit popularity on fashion-oriented websites to supervise the scoring element. The scoring element may be a multimodal multiinstance deep learning system that evaluates instance aesthetics and set compatibility simultaneously. In order to train and evaluate the proposed composition system, we tend to have collected a giant-scale fashion outfit dataset with 195K outfits and 368K fashion items from Polyvore. Although the fashion outfit scoring and composition is rather difficult, we have a tendency to have achieved an AUC of eighty fivepercent for the scoring component, and an accuracy of seventy sevenp.c for a constrained composition task.