Skew Estimation by Improved Boundary Growing for Text Documents in South Indian Languages

摘要

Estimating the inclination of lines in skewed documents made up of texts in south Indian languages (Kannada, Telugu, Tamil and Malayalam) is not as straight forward as computing the skew of text documents in English. This is due to additional modifier-characters, which get plugged in as bottom fixes or top fixes, or as extensions, that remain as disconnected protrusions of a main character. Under such circumstances direct application of Boundary Growing (BG) method would fail to perform accurately, hence we have proposed a corrective step employing Nearest Neighbor Clustering (NNC). BG and NNC jointly derive the coordinates to be input into moments computation to estimate the angle of inclination. The new model is tested on varieties of documents containing noisy texts, mixed with pictures, text in different resolutions which are composed in south Indian languages Kannada, Telugu, Tamil and Malayalam. For the purpose of contrasting, texts in English are also considered.