Features for Word Spotting in Historical Manuscripts

For the transition from traditional to digital libraries, the large number of handwritten manuscripts that exist pose a great challenge. Easy access to such collections requires an index, which is currently created manually at great cost. Because automatic handwriting recognizers fail on histori-cal manuscripts, the word spotting technique has been de-veloped: the words in a collection are matched as images and grouped into clusters which contain all instances of the same word. By annotating “interesting ” clusters, an index that links words to the locations where they occur can be built automatically. Due to the noise in historical documents, selecting the right features for matching words is crucial. We analyzed a range of features suitable for matching words using dy-namic time warping (DTW), which aligns and compares sets of features extracted from two images. Each feature’s individual performance was measured on a test set. With an average precision of 72%, a combination of features outper-forms competing techniques in speed and precision. 1