SAT: Self-adaptive training for fashion compatibility prediction

TL;DR

This paper introduces a self-adaptive training method with a novel triplet loss for improving fashion compatibility prediction, especially for hard-to-classify outfits, demonstrating effectiveness on public datasets.

Contribution

The paper proposes a new self-adaptive triplet loss and a difficulty score to better learn hard fashion items, enhancing compatibility prediction accuracy.

Findings

Effective in predicting fashion compatibility on Polyvore datasets

Improves learning of hard-to-classify outfits

Can be extended to other similarity networks

Abstract

This paper presents a self-adaptive training (SAT) model for fashion compatibility prediction. It focuses on the learning of some hard items, such as those that share similar color, texture, and pattern features but are considered incompatible due to the aesthetics or temporal shifts. Specifically, we first design a method to define hard outfits and a difficulty score (DS) is defined and assigned to each outfit based on the difficulty in recommending an item for it. Then, we propose a self-adaptive triplet loss (SATL), where the DS of the outfit is considered. Finally, we propose a very simple conditional similarity network combining the proposed SATL to achieve the learning of hard items in the fashion compatibility prediction. Experiments on the publicly available Polyvore Outfits and Polyvore Outfits-D datasets demonstrate our SAT's effectiveness in fashion compatibility prediction. Besides, our SATL can be easily extended to other conditional similarity networks to improve their performance.