Unitail: Detecting, Reading, and Matching in Retail Scene

TL;DR

Unitail is a comprehensive large-scale dataset and benchmark designed to advance detection, reading, and matching of retail products, facilitating improved computer vision applications in store environments.

Contribution

The paper introduces Unitail, a new large-scale dataset with annotations for detection, OCR, and matching tasks tailored for retail scenes, along with customized algorithms and benchmarks.

Findings

The dataset contains 1.8 million annotated instances for detection.

The OCR dataset includes 1454 product categories and 30,000 text regions.

The customized detector and OCR-based matching demonstrate effective performance.

Abstract

To make full use of computer vision technology in stores, it is required to consider the actual needs that fit the characteristics of the retail scene. Pursuing this goal, we introduce the United Retail Datasets (Unitail), a large-scale benchmark of basic visual tasks on products that challenges algorithms for detecting, reading, and matching. With 1.8M quadrilateral-shaped instances annotated, the Unitail offers a detection dataset to align product appearance better. Furthermore, it provides a gallery-style OCR dataset containing 1454 product categories, 30k text regions, and 21k transcriptions to enable robust reading on products and motivate enhanced product matching. Besides benchmarking the datasets using various state-of-the-arts, we customize a new detector for product detection and provide a simple OCR-based matching solution that verifies its effectiveness.