DeepFashion2: A Versatile Benchmark for Detection, Pose Estimation, Segmentation and Re-Identification of Clothing Images

TL;DR

DeepFashion2 introduces a comprehensive benchmark dataset with rich annotations for clothing detection, pose estimation, segmentation, and re-identification, addressing limitations of previous datasets and supporting advanced fashion image understanding.

Contribution

It provides a large, richly annotated dataset and a strong baseline model for multiple fashion image analysis tasks, bridging gaps in real-world scenario representation.

Findings

DeepFashion2 contains 801K clothing items with detailed annotations.

The dataset includes 873K commercial-consumer clothing pairs.

Extensive evaluations demonstrate the effectiveness of the proposed baseline.

Abstract

Understanding fashion images has been advanced by benchmarks with rich annotations such as DeepFashion, whose labels include clothing categories, landmarks, and consumer-commercial image pairs. However, DeepFashion has nonnegligible issues such as single clothing-item per image, sparse landmarks (4~8 only), and no per-pixel masks, making it had significant gap from real-world scenarios. We fill in the gap by presenting DeepFashion2 to address these issues. It is a versatile benchmark of four tasks including clothes detection, pose estimation, segmentation, and retrieval. It has 801K clothing items where each item has rich annotations such as style, scale, viewpoint, occlusion, bounding box, dense landmarks and masks. There are also 873K Commercial-Consumer clothes pairs. A strong baseline is proposed, called Match R-CNN, which builds upon Mask R-CNN to solve the above four tasks in an end-to-end manner. Extensive evaluations are conducted with different criterions in DeepFashion2.