3D text recognition is an area that develops fast in the past years due to the great improvement in the computational capabilities of hand devises and computer systems, and the increase in the imaging quality of digital cameras.
3D text recognition using digital cameras is different from traditional 2D text recognition using scanners in many ways. 3D effects have to be considered and eliminated before the performing of 3D text recognition.Hence, new methodologies have to be developed for 3D text recognition.
This work is devoted to the preprocessing part of 3D text recognition by performing image enhancement and binarization for further processing and recognition.
Experiments were conducted on various 3D text images. The results verify the validity and efficiency of the proposed method in the binarization of 3D text images. Moreover, the recognition rate can be enhanced by using the proposed binarization method before the recognition.