|
|
|
 |
|
|
|
|
| Features | Download | Buy | News | Support |
 |
|
|
|
|
The following examples illustrate various application categories that SmartOCR can handle. The benchmarks were performed on a 1.7 GHz Pentium 4 with 512 MB RAM running Windows 2000. All the target images are 640x480 unsigned 8-bit. In all cases, the entire target image is searched.
The same font was used across all the application categories to minimize complex interaction of conditions, which may confuse the meaning of the benchmarks. All the characters were prepared with Microsoft Word and printed on a laser printer. The font models were prepared with Arial size 16. All the images were acquired using a standard RS-170 camera with a Coreco Imaging Bandit II frame grabber, with no special lighting other than desk lamps and overhead flourescent lighting. The size of each character used for training is 20x30 pixels.
| Category | Font | Image | Parameters |
Speed (ms) | Char/sec |
|
Simple: clean image, single target, low noise, no rotation, no scale changes, some characters may be rotationally ambiguous (e.g. 6 and 9) |
Digits only,
binary
|
| All default values | 1.65 | 6000 |
|
Rotation: Characters in image rotated by about 30° relative to trained font. |
Letters and digits, binary, 36 characters 
|
| All default values | 3.28 | 9000 |
|
Scale: Characters in image are 3 times the size of the font model. The scale change was introduced by selecting a larger font size in the word processor. |
Digits only,
binary
|
| All default values | 2.1 | 3000 |
|
Damage: Characters partially erased on wrinkled paper. No rotation or scale changes. |
Digits only, grayscale
|
| Method = area | 18.6 | 500 |
|
Occlusion: Characters partially occluded by strips of paper. No rotation or scale changes. |
Digits only, grayscale
|
| Method = area | 16.9 | 600 |
