Visual Saliency and Biological Inspired Text Detection Export

@article{citeulike:5455853, abstract = {The outline of the thesis is as follows: the next chapter provides the different methods on automatic text detection including why Optical Character Recognition (OCR) techniques are not applicable as text detectors on urban environment scenes in Section 2.1, a Related Work in Section 2.2, Feature Selection in Section 2.3, an introduction to the used statistical machine learning technique, the Support Vector Machine (SVM) in Section 2.4. Chapter 3 gives an introduction to visual attention and the Saliency Model by Itti and Koch [2001] which is used in this thesis as the computational model of bottom-up visual attention. Chapter 4 describes the combination of a text detection algorithm (SalTd) with the standard Saliency Model. Implementation details of the SalTd algorithm based on Gabor features and SVM is provided in Chapter 5. The Results are discussed in Chapter 6. Section 6.1 describes the results of the text detection algorithm. Section 6.2 gives the comparison of the SalTd algorithm implemented within this thesis with an AdaBoost (Adaptive Boosting) based state-of-the-art text detection algorithm. The last section of the results chapter provides the comparison of the standard Saliency Model com- bined with the text detection algorithm and the standard Saliency Model's eye movement predictions. A summary and suggestions for further work is given in Chapter 7.}, author = {Konuskan, Fatma}, citeulike-article-id = {5455853}, comment = {* cites Mancas embedded OCR paper}, keywords = {ocr, text-localization}, posted-at = {2009-08-18 02:29:32}, priority = {2}, title = {Visual Saliency and Biological Inspired Text Detection}, year = {2008} }

TY - JOUR ID - citeulike:5455853 L3 - citeulike-article-id:5455853 N2 - The outline of the thesis is as follows: the next chapter provides the different methods on automatic text detection including why Optical Character Recognition (OCR) techniques are not applicable as text detectors on urban environment scenes in Section 2.1, a Related Work in Section 2.2, Feature Selection in Section 2.3, an introduction to the used statistical machine learning technique, the Support Vector Machine (SVM) in Section 2.4. Chapter 3 gives an introduction to visual attention and the Saliency Model by Itti and Koch [2001] which is used in this thesis as the computational model of bottom-up visual attention. Chapter 4 describes the combination of a text detection algorithm (SalTd) with the standard Saliency Model. Implementation details of the SalTd algorithm based on Gabor features and SVM is provided in Chapter 5. The Results are discussed in Chapter 6. Section 6.1 describes the results of the text detection algorithm. Section 6.2 gives the comparison of the SalTd algorithm implemented within this thesis with an AdaBoost (Adaptive Boosting) based state-of-the-art text detection algorithm. The last section of the results chapter provides the comparison of the standard Saliency Model com- bined with the text detection algorithm and the standard Saliency Model’s eye movement predictions. A summary and suggestions for further work is given in Chapter 7. TI - Visual Saliency and Biological Inspired Text Detection KW - ocr KW - text-localization N1 - * cites Mancas embedded OCR paper AU - Konuskan, Fatma PY - 2008/// ER -