Segmented Characters v0.1

Segmented characters, is a dataset that can be used for training/testing optical character recognition(OCR) systems. Dataset includes numbers between 0-9 and uppercase letters. Because data is gathered from the real world, data are noisy.
This data set is pretty incomplete so, you can use it to expand your current dataset.
I will be glad if you inform me about problems at b.evrim*AT* gmail *DOT* com

Examples from dataset:

Which character how many?

Character count 0 100 1 114 2 131 3 92 4 111 5 92 6 102 7 87 8 124 9 105 A 191 B 70 C 69 D 4 E 6

Character count F 1 G 3 H 31 K 92 L 6 M 58 N 5 O 4 P 47 T 26 V 1 X 144 Y 4 Z 2 . .

total: 1821

Download:
Segmented Characters v0.1

---

OTHER DATASETS:

• Handwriting characters dataset: consists of 52152 characters

• Robust character set: 3 sets consists of 854, 6185 and 5430 characters