Segmentation-free word spotting with exemplar SVMs
作者：
Highlights：
• An unsupervised segmentation-free word spotting method is proposed.
• A sliding window approach is used to locate document regions similar to the query image.
• Exemplar SVM is used to improve the query representation.
• PQ compression allows one to deal with large collections of documents with a minimum loss in accuracy.
• Reranking and query expansion can be applied in word spotting to improve performance.
摘要
Highlights•An unsupervised segmentation-free word spotting method is proposed.•A sliding window approach is used to locate document regions similar to the query image.•Exemplar SVM is used to improve the query representation.•PQ compression allows one to deal with large collections of documents with a minimum loss in accuracy.•Reranking and query expansion can be applied in word spotting to improve performance.
论文关键词：Word spotting,Segmentation-free,Unsupervised learning,Reranking,Query expansion,Compression
论文评审过程：Received 2 July 2013, Revised 4 June 2014, Accepted 6 June 2014, Available online 17 June 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.06.005
