Research Article
Practical Recognition System for Text Printed on Clear Reflected Material
Table 11
Comparison with other systems.
| Research | Language | Method of segmentation | Rate | Number of images | Year |
| [44] | English | Color | 0.9 | 68 | 2006 | [45] | English | Edge algorithm | 0.75 | Not reported | 2007 | [46] | English | Altering | 0.925 | Not reported | 2001 | [47] | English | Convolution | 0.82 | 1615 | 2008 | [48] | Malaysian | Morphological processes | 0.95 | 589 | 2008 | [11] | Japanese | Neural network | 0.78 | 100 | 2005 | [12] | English | Neighboring | 0.96 | 30 | 2008 | [49] | Korean | Niblack threshold | 0.88 | 33 | 2008 | [50] | China | GMMs | 0.85 | 1000 | 2004 | [13] | Arabic | SelCT | 0.88 | Arabic news story | 2007 | [51] | English | Fuzzy clustering | 0.97 | 25 | 2006 | Proposed Slab | English | Proposed method—Slab | 0.99 | 10 different sets (20 each set) | 2010 | Proposed Dasani | English | Proposed method—Dasani | 0.985 | 50 different sets | 2010 | Proposed Ozarka | English | Proposed method—Ozarka | 0.97 | 30 different sets | 2010 |
| Proposed average | | 0.98 | | |
|
|