With the rapid development of electronic commerce, the way of shopping has experienced a revolutionary evolution. To fully meet customers’ massive and diverse online shopping needs with quick response, retailing AI system needs to automatically recognize products from images and videos at the stock-keeping unit (SKU) level with high accuracy. However, product recognition is still a challenging task, since many of SKU-level products are fine-grained and visually similar by a rough glimpse. Although there are already some products benchmarks available, these datasets are either too small (limited number of products) or noisy-labeled (lack of human labeling).
We construct a human-labeled products image dataset named “Products-10k”, which is so far the largest production recognition dataset containing 10,000 products frequently bought by online customers in JD.com, covering a full spectrum of categories including Fashion, 3C, food, healthcare, household commodities, etc.. Moreover, large-scale product labels are organized as a graph to indicate the complex hierarchy and inter-dependency among products.
Based on this dataset, we organize the 1st Challenge on SKU-level production recognition. The final results will be announced at ICPR2020, and the winner will be invited to present their approaches at the workshop. We encourage engineers and researchers from the pattern recognition community to develop novel algorithms for this practical and challenging task.
Due to the difficulties imposed by the global coronavirus epidemic, we decided to delay the Challenge Lauch Date from May to August this year.
Competition URL https://www.kaggle.com/c/products-10k
|Challenge Launch Data||August 20, 2020|
|Challenge Submissions deadline||Sepetember 30, 2020|
|Challenge Award Notification||October 7, 2020|
Please cite the following paper if you use our dataset.
Yalong Bai, Yuxiang Chen, Wei Yu, Linfang Wang, Wei Zhang. "Products-10K: A Large-scale Product Recognition Dataset". [arXiv]