Abstract | ||
---|---|---|
With the high-growing of e-commerce, more and more users have changed to buy from websites rather than in stores. To deal with mass products, the traditional text-based product search has become incompetent to meet use’s requirement. In this paper, we explore deep learning with convolutional neural networks (CNN) to resolve query’s classification, and propose an efficient approach for product image search. For a query image, we first train a CNN model of a large database containing various product images to discriminate the query’s category. Then we search similar products from the established category and utilize these visual results to parse the query with attribute. Finally we use the extracted attribute tags to finish the textual re-ranking and obtain the most relevant retrieved product list. Experimental evaluation shows that our approach significantly outperforms state of art in product image search. |
Year | Venue | Field |
---|---|---|
2016 | PCM | Pattern recognition,Information retrieval,Ranking,Computer science,Convolutional neural network,Artificial intelligence,Parsing,Deep learning |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
14 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xin Zhou | 1 | 25 | 13.16 |
Yuqi Zhang | 2 | 4 | 7.17 |
Xiuxiu Bai | 3 | 1 | 2.03 |
Jihua Zhu | 4 | 51 | 6.90 |
Li Zhu | 5 | 32 | 9.37 |
Xueming Qian | 6 | 1052 | 70.70 |