OCR Language Models with Custom Vocabularies
This addresses the issue of poor OCR performance for specialized documents like checks and medical prescriptions, though it is incremental as it builds on existing language model and OCR methods.
The paper tackles the problem of optical character recognition (OCR) accuracy in specialized domains by introducing an algorithm to generate domain-specific language models at runtime and a modified CTC beam search decoder, resulting in a substantial reduction in word error rate.
Language models are useful adjuncts to optical models for producing accurate optical character recognition (OCR) results. One factor which limits the power of language models in this context is the existence of many specialized domains with language statistics very different from those implied by a general language model - think of checks, medical prescriptions, and many other specialized document classes. This paper introduces an algorithm for efficiently generating and attaching a domain specific word based language model at run time to a general language model in an OCR system. In order to best use this model the paper also introduces a modified CTC beam search decoder which effectively allows hypotheses to remain in contention based on possible future completion of vocabulary words. The result is a substantial reduction in word error rate in recognizing material from specialized domains.