Abstract:Instance-level image retrieval in fashion is a challenging issue owing to its increasing importance in real-scenario visual fashion search. Cross-domain fashion retrieval aims to match the unconstrained customer images as queries for photographs provided by retailers; however, it is a difficult task due to a wide range of consumer-to-shop (C2S) domain discrepancies and also considering that clothing image is vulnerable to various non-rigid deformations. To this end, we propose a novel multi-scale and multi-granularity feature learning network (MMFL-Net), which can jointly learn global-local aggregation feature representations of clothing images in a unified framework, aiming to train a cross-domain model for C2S fashion visual similarity. First, a new semantic-spatial feature fusion part is designed to bridge the semantic-spatial gap by applying top-down and bottom-up bidirectional multi-scale feature fusion. Next, a multi-branch deep network architecture is introduced to capture global salient, part-informed, and local detailed information, and extracting robust and discrimination feature embedding by integrating the similarity learning of coarse-to-fine embedding with the multiple granularities. Finally, the improved trihard loss, center loss, and multi-task classification loss are adopted for our MMFL-Net, which can jointly optimize intra-class and inter-class distance and thus explicitly improve intra-class compactness and inter-class discriminability between its visual representations for feature learning. Furthermore, our proposed model also combines the multi-task attribute recognition and classification module with multi-label semantic attributes and product ID labels. Experimental results demonstrate that our proposed MMFL-Net achieves significant improvement over the state-of-the-art methods on the two datasets, DeepFashion-C2S and Street2Shop.

Fashion sub-categories and attributes prediction model using deep learning

A Deep-Learning-Based Fashion Attributes Detection Model

A knowledge-sharing semi-supervised approach for fashion clothes classification and attribute prediction

Personalized Clothing Prediction Algorithm Based on Multi-modal Feature Fusion

Fashion Analysis With A Subordinate Attribute Classification Network

Towards Better Understanding the Clothing Fashion Styles: A Multimodal Deep Learning Approach

Multiple-Clothing Detection and Fashion Landmark Estimation Using a Single-Stage Detector

Deep Learning for Demand Forecasting in the Fashion and Apparel Retail Industry

Deep Learning Image Classification for Fashion Design

Clothes Keypoints Localization and Attribute Recognition Via Prior Knowledge.

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

A Federated Approach for Fine-Grained Classification of Fashion Apparel

MMFashion: An Open-Source Toolbox for Visual Fashion Analysis

FashionNet: Personalized Outfit Recommendation with Deep Neural Network

Unsupervised Deep Clustering for Fashion Images.

FFENet: Frequency-Spatial Feature Enhancement Network for Clothing Classification

Describe Fashion Products via Local Sparse Self-Attention Mechanism and Attribute-based Re-sampling Strategy

MMFL-Net: Multi-scale and Multi-granularity Feature Learning for Cross-domain Fashion Retrieval

Multideep Feature Fusion Algorithm for Clothing Style Recognition

A deep learning system for predicting size and fit in fashion e-commerce

Efficient Clothing Fashion Prediction using Machine Learning