$extrastylesheet
Functions | |
template<typename I > | |
mln::trait::ch_value< I, bool > ::ret | scribo::binarization::kim (const Image< I > &input, unsigned window_size, double k) |
template<typename I > | |
mln::trait::ch_value< I, bool > ::ret | scribo::binarization::kim (const Image< I > &input, unsigned window_size) |
Kim's Binarization implementations.
This algorithms performs a first rough binarization on the input (here we use Sauvola's method). Then text lines are roughly detected by grouping connected components. For each text lines, character thickness and x height is computed. Finally, each lines is binarized again, using text features as parameters for adjusting the final threshold.
This algorithms considers that global and local statistics should be used to compute a threshold in text areas.
This is an improvement of Sauvola's method.
This implementation is based on the paper "Multi-Window Binarization of Camera Image for Document Recognition", In-Jung Kim, Proceedings of the 9th International Workshop on Frontiers in Handwriting Recognition (IWFHR-9 2004)
mln::trait::ch_value< I , bool >::ret scribo::binarization::kim | ( | const Image< I > & | input, |
unsigned | window_size, | ||
double | k | ||
) |
Kim's binarization.
This algorithms performs a first rough binarization on the input (here we use Sauvola's method). Then text lines are roughly detected by grouping connected components. For each text lines, character thickness and x height is computed. Finally, each lines is binarized again, using text features as parameters for adjusting the final threshold.This algorithms considers that global and local statistics should be used to compute a threshold in text areas.This is an improvement of Sauvola's method.This implementation is based on the paper "Multi-Window Binarization of Camera Image for Document Recognition", In-Jung Kim, Proceedings of the 9th International Workshop on Frontiers in Handwriting Recognition (IWFHR-9 2004)
[in] | input | A gray-level image. |
[in] | window_size | the window size to be used for the first binarization. |
[in] | k | Sauvola's formula parameter. |
mln::trait::ch_value< I , bool >::ret scribo::binarization::kim | ( | const Image< I > & | input, |
unsigned | window_size | ||
) |
This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts. k is set to SCRIBO_DEFAULT_SAUVOLA_K.