Timeline for OCR of low-resolution text from screenshots
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 11, 2012 at 18:22 | vote | accept | mpenkov | ||
| Jan 2, 2012 at 5:19 | comment | added | rsaxvc | Slower than subtraction, but you can search an image for all instances of a certain cascade at once. | |
| Jan 2, 2012 at 5:17 | comment | added | rsaxvc | You'll have to generate cascades, which is a pretty time-consuming process(but also stupidly parallelizeable). It also requires a bunch of input data( I would use the numbers in every font you have on a desktop). | |
| Jan 2, 2012 at 5:03 | comment | added | mpenkov | Thanks! I'll definitely have a look at Haar cascades. How efficient do you think it will be compared to simple image subtraction? I expect it to be slower. However, if it's 5 times slower but replaces 10 image checks for the same effectiveness, then it would definitely be worth it. | |
| Jan 2, 2012 at 4:17 | history | answered | rsaxvc | CC BY-SA 3.0 |