Method and apparatus for Arabic and Farsi font recognition

A method of font recognition comprising:

• calculating a plurality of feature values for a sample text, wherein the feature values correspond to a plurality of font features, and the plurality of font features includes Log-Gabor orientation-scale-variance features and Log-Gabor orientation-scale-mean features for the sample text;
• determining a Euclidian distance between the plurality of the feature values for the sample text and respective model feature values for each of a plurality of predefined fonts, wherein the Euclidian distance for the ith predefined font is given by
• D
• i
• =
• ∑
• j
• =
• 1
• n
• ⁢
• ⁢
• (
• T
• ij
• -
• V
• j
• )
• 2
• ,
• where n is the number of feature values, Tij is an element corresponding to the ith row and jth column of a matrix of the model values, and Vj is a jth element a vector containing the plurality of the feature values for the sample text;
• and
• signaling that the font of the sample text is the font from the plurality of predefined fonts corresponding to the smallest Euclidian distance.

An apparatus for font recognition, the apparatus comprising:

• circuitry configured to calculate a plurality of feature values for a sample text, wherein the feature values correspond to a plurality of font features, and the plurality of font features includes Log-Gabor orientation-scale-variance features and Log-Gabor orientation-scale-mean features for the sample text;
• circuitry configured to calculate a Euclidian distance between the plurality of the feature values for the sample text and respective model feature values for each of a plurality of predefined fonts, wherein the Euclidian distance for the ith predefined font is given by
• D
• i
• =
• ∑
• j
• =
• 1
• n
• ⁢
• ⁢
• (
• T
• ij
• -
• V
• j
• )
• 2
• ,
• where n is the number of feature values, Tij is an element corresponding to the ith row and jth column of a matrix of the model values, and Vj is a jth element a vector containing the plurality of the feature values for the sample text;
• and
• circuitry configured to determine the smallest Euclidian distance, and to signal that the font of the sample text is the font from the plurality of predefined fonts corresponding to the smallest Euclidian distance.