This study aims to extract hidden words from continuous text (obtained by removing spaces from original text) by detecting word boundaries and performing text segmentation.
Viterbi Algorithm is used on a Unigram model of word probabilities to retrieve the original text by inserting spaces at correct locations with the aim of maximizing the cosine similarity of the resultant text and the original (formatted) text. The analysis of the results produced by comparison confirms the efficiency of this approach to solve the problem.