FoodLogoDet-1500: A Dataset for Large-Scale Food Logo Detection via Multi-Scale Feature Decoupling Network

TL;DR

This paper introduces FoodLogoDet-1500, a large-scale food logo dataset with 1,500 categories and 100,000 images, and proposes MFDNet, a novel detection network that improves food logo recognition by decoupling classification and regression tasks.

Contribution

The paper presents the first large-scale food logo dataset and a new detection method, MFDNet, that enhances classification accuracy through feature decoupling and balanced multi-scale features.

Findings

FoodLogoDet-1500 contains 1,500 categories and 100,000 images.

MFDNet outperforms existing methods on food logo detection.

The proposed method effectively distinguishes similar food logo categories.

Abstract

Food logo detection plays an important role in the multimedia for its wide real-world applications, such as food recommendation of the self-service shop and infringement detection on e-commerce platforms. A large-scale food logo dataset is urgently needed for developing advanced food logo detection algorithms. However, there are no available food logo datasets with food brand information. To support efforts towards food logo detection, we introduce the dataset FoodLogoDet-1500, a new large-scale publicly available food logo dataset, which has 1,500 categories, about 100,000 images and about 150,000 manually annotated food logo objects. We describe the collection and annotation process of FoodLogoDet-1500, analyze its scale and diversity, and compare it with other logo datasets. To the best of our knowledge, FoodLogoDet-1500 is the first largest publicly available high-quality dataset for food logo detection. The challenge of food logo detection lies in the large-scale categories and similarities between food logo categories. For that, we propose a novel food logo detection method Multi-scale Feature Decoupling Network (MFDNet), which decouples classification and regression into two branches and focuses on the classification branch to solve the problem of distinguishing multiple food logo categories. Specifically, we introduce the feature offset module, which utilizes the deformation-learning for optimal classification offset and can effectively obtain the most representative features of classification in detection. In addition, we adopt a balanced feature pyramid in MFDNet, which pays attention to global information, balances the multi-scale feature maps, and enhances feature extraction capability. Comprehensive experiments on FoodLogoDet-1500 and other two benchmark logo datasets demonstrate the effectiveness of the proposed method. The FoodLogoDet-1500 can be found at this https URL.