LogoDet-3K: A Large-Scale Image Dataset for Logo Detection

December 18, 2023 ยท View on GitHub

LogoDet-3K-Dataset

example

LogoDet-3K Dataset Description

In this work, we introduce LogoDet-3K, the largest logo detection dataset with full annotation, which has 3,000 logo categories, about 200,000 manually annotated logo objects and 158,652 images. LogoDet-3K creates a more challenging benchmark for logo detection, for its higher comprehensive coverage and wider variety in both logo categories and annotated objects compared with existing datasets. The statistic comparison of 9 super categories from LogoDet-3K is shown as follows,

Root CategorySub-CategoryImagesObjects
Food93253,35064,276
Clothes60431,26637,601
Necessities43224,82230,643
Others37115,51320,016
Electronic2249,67512,139
Transportation21310,44512,791
Leisure1115,6856,573
Sports663,9455,041
Medical473,9455,185
Total3,000158,652194,261

LogoDet-3K Dataset Detailed Statistics

The figure showns the detailed statistics of LogoDet-3K about Image and object distribution in per category, and the number of objects in per image and object size in per image. example

The sorted distribution of images for each logo in LogoDet-3K is shown the figure. We can see that imbalanced distribution across different logo categories are one characteristic of LogoDet-3K, posing a challenge for effective logo detection with few samples. example

The figure shows the distributions of categories, images and objects from LogoDet-3K on super-classes. example

Server link: 123.57.42.89/Dataset_ict/LogoDet-3K.zip

Baidu Drive link: https://pan.baidu.com/share/init?surl=ECJ4zOurhMzBgDZ11ekxQA password:1234

Google Drive link: https://drive.google.com/file/d/1GOYqRy7yoBVAt9Ih375NWtoZZTXN6OET/view?usp=sharing