The Open Brands Dataset: Unified brand detection and recognition at scale

TL;DR

This paper introduces the Open Brands dataset, a large-scale, richly annotated benchmark for brand detection and recognition, and proposes a new model called Brand Net that achieves state-of-the-art performance.

Contribution

It provides the largest comprehensive dataset for brand recognition and detection, along with a novel model tailored for this task, addressing limitations of previous small-scale datasets.

Findings

Open Brands contains over 1.4 million images with extensive annotations.

Brand Net achieves state-of-the-art mAP on the dataset.

Performance improves with increased training data.

Abstract

Intellectual property protection(IPP) have received more and more attention recently due to the development of the global e-commerce platforms. brand recognition plays a significant role in IPP. Recent studies for brand recognition and detection are based on small-scale datasets that are not comprehensive enough when exploring emerging deep learning techniques. Moreover, it is challenging to evaluate the true performance of brand detection methods in realistic and open scenes. In order to tackle these problems, we first define the special issues of brand detection and recognition compared with generic object detection. Second, a novel brands benchmark called "Open Brands" is established. The dataset contains 1,437,812 images which have brands and 50,000 images without any brand. The part with brands in Open Brands contains 3,113,828 instances annotated in 3 dimensions: 4 types, 559 brands and 1216 logos. To the best of our knowledge, it is the largest dataset for brand detection and recognition with rich annotations. We provide in-depth comprehensive statistics about the dataset, validate the quality of the annotations and study how the performance of many modern models evolves with an increasing amount of training data. Third, we design a network called "Brand Net" to handle brand recognition. Brand Net gets state-of-art mAP on Open Brand compared with existing detection methods.