Handwritten-Word Spotting Using Biologically Inspired Features

Abstract

For quick access to new handwritten collections, current handwriting recognition methods are too cumbersome. They cannot deal with the lack of labeled data and would require extensive laboratory training for each individual script, style, language and collection. We propose a biologically inspired whole-word recognition method which is used to incrementally elicit word labels in a live, web-based annotation system, named Monk. Since human labor should be minimized given the massive amount of image data, it becomes important to rely on robust perceptual mechanisms in the machine. Recent computational models of the neuro-physiology of vision are applied to isolated word classification. A primate cortex-like mechanism allows to classify text-images that have a low frequency of occurrence. Typically these images are the most difficult to retrieve and often contain named entities and are regarded as the most important to people. Usually standard pattern-recognition technology cannot deal with these text-images if there are not enough labeled instances. The results of this retrieval system are compared to normalized word-image matching and appear to be very promising.

Towards Style-Based Dating of Historical Documents

“…The collected database and modeling are not only relevant to the problem of dating, but also help to develop character and word models for classification and retrieval [4], [6].…”

“…However, there exists virtually no literature about dating historical documents or books. The emergence of historical handwriting collections, such as the Monk system [4] which contains more than 50k page images, has enabled progress on novel historical document dating methodology using pattern recognition and machine learning.…”

Merging Grid Technologies

“…* target processing 19 -the main web tool to process the data in Astro-WISE. This web interface allows users to go through pre-defined…”

“…By combining brute force 46 http://www.rug.nl/ai (correlator) matching with traditional machine learning using Hidden-Markov Models and Support Vector Machines, it was proved that continuous web-based transcription of text, concurrent with nightly retraining computation sessions allows to open up access to large collections of scans of handwritten text. The description of the methods behind this work can be found in details in [19] and [20].…”

Segmental K-Means Learning with Mixture Distribution for HMM Based Handwriting Recognition

“…Therefore, algorithms are needed which can handle grey-scale. This is important in situations with low-quality images, such as in the Monk system [4] for historical manuscript retrieval.…”