WEBVTT

00:00:00.080 --> 00:00:06.640
you know it's pretty easy to take words on your computer screen and put them on

00:00:04.000 --> 00:00:10.880
a physical sheet of paper just click print and unless you've forgotten to

00:00:08.240 --> 00:00:16.560
fork out an extortion level amount of money for a new cartridge you'll have

00:00:13.120 --> 00:00:18.560
fresh warm satisfying documents just a

00:00:16.560 --> 00:00:22.960
few moments later but going in the opposite direction scanning dead tree

00:00:20.800 --> 00:00:27.359
information into your pc is actually quite a bit trickier i mean sure flatbed

00:00:25.519 --> 00:00:30.720
scanners aren't all that difficult to operate per se

00:00:28.800 --> 00:00:35.280
but many of them are basically just taking a picture of the document and

00:00:32.880 --> 00:00:39.760
saving it onto your pc meaning not only will it probably not look very crisp due

00:00:37.760 --> 00:00:44.640
to file compression and little bits of dust in your scanner but you can't edit

00:00:42.480 --> 00:00:48.239
a clean copy of your document in your favorite word processor because the

00:00:46.160 --> 00:00:50.960
scanner won't recognize each individual character

00:00:49.600 --> 00:00:55.920
fortunately there are a number of devices out there that enable optical

00:00:52.960 --> 00:01:00.480
character recognition or ocr where each character on a page is scanned

00:00:57.920 --> 00:01:05.920
individually so your papers are uploaded as actual text documents instead of

00:01:02.800 --> 00:01:08.479
messy jpegs but how exactly does that

00:01:05.920 --> 00:01:12.479
work and is one kind of optical scanner better than another well because the

00:01:10.240 --> 00:01:16.560
whole concept of translating text into electronic signal is pretty broad there

00:01:15.439 --> 00:01:21.840
have been lots of different implementations of ocr over the years in

00:01:19.280 --> 00:01:27.360
fact one of the earliest electric ocr devices the optophone was invented all

00:01:24.880 --> 00:01:33.200
the way back in 1914 this bizarre looking contraption relied

00:01:29.840 --> 00:01:35.439
on the special behavior of selenium

00:01:33.200 --> 00:01:39.920
which conducts electricity differently in light and darkness

00:01:37.520 --> 00:01:45.600
as it scanned the words on a page the optophone distinguished between the dark

00:01:42.320 --> 00:01:47.759
ink of text and lighter blank spaces

00:01:45.600 --> 00:01:53.280
generating tones that corresponded to different letters making it possible for

00:01:49.600 --> 00:01:56.000
blind people to read with some practice

00:01:53.280 --> 00:01:59.920
later in 1931 a machine was developed that could convert printed text to

00:01:58.079 --> 00:02:04.399
telegraph code one of the first technologies to translate printed

00:02:01.759 --> 00:02:09.840
characters to electrical impulses rather than sounds but it wasn't until the

00:02:06.520 --> 00:02:12.480
1960s and 70s that ocr began to take a

00:02:09.840 --> 00:02:17.440
more familiar modern form with postal services using ocr to read addresses and

00:02:15.680 --> 00:02:21.440
software that could recognize many different fonts

00:02:18.800 --> 00:02:25.760
so back to present day when you scan a document how exactly does the software

00:02:23.599 --> 00:02:30.720
know what it's looking at well the first step is to cut out artifacts so your ocr

00:02:28.319 --> 00:02:35.680
program can concentrate on the text and nothing else so it attempts to remove

00:02:32.879 --> 00:02:41.200
dust and other various graphics align the text properly and convert any colors

00:02:38.239 --> 00:02:45.440
or shades of grey in the image to black and white only making the words

00:02:42.959 --> 00:02:50.000
themselves easier to recognize the next step is to figure out which characters

00:02:47.120 --> 00:02:54.800
are on the page simpler forms of ocr compare each scanned letter pixel by

00:02:52.080 --> 00:03:00.400
pixel to a known database of fonts and decide on the closest match smarter ocr

00:02:58.080 --> 00:03:05.360
however takes this step farther by breaking down each character down to

00:03:02.640 --> 00:03:09.840
constituent elements like curves and corners and looking for matching

00:03:07.280 --> 00:03:13.120
physical features and actual letters you can think of the differences between

00:03:11.280 --> 00:03:17.040
these two approaches similarly to the difference between raster and vector

00:03:15.040 --> 00:03:22.080
images which you can learn more about up here ocr software can also make use of a

00:03:20.319 --> 00:03:26.319
dictionary so it won't accidentally spit out nonsense words due to inaccurate

00:03:24.400 --> 00:03:31.519
scanning for example if your scanner sees this but it can't quite tell

00:03:28.879 --> 00:03:36.560
whether the middle letter is an o or an a it can check its own dictionary to

00:03:33.760 --> 00:03:42.080
decide that the word is actually dog and not dag giving ocr software situational

00:03:40.000 --> 00:03:47.760
information can further cut down on errors such as telling it to only try to

00:03:44.959 --> 00:03:51.519
match numbers if it's reading zip codes on an envelope

00:03:49.200 --> 00:03:56.239
even with these tricks however ocr obviously is not perfect which you've

00:03:54.159 --> 00:04:00.640
probably seen for yourself if you've ever used it but with greater

00:03:58.879 --> 00:04:04.799
processing power and machine learning techniques that allow software to

00:04:02.400 --> 00:04:09.439
recognize more subtle patterns over time ocr has become versatile enough to

00:04:07.200 --> 00:04:13.360
recognize harder to read typefaces inconsistently printed material and even

00:04:12.000 --> 00:04:17.919
handwriting and free ocr cloud processing services

00:04:15.760 --> 00:04:22.240
like google drive which has a lot more machine learning capability than your

00:04:19.519 --> 00:04:27.600
home pc for which i hope are fairly obvious reasons have made ocr more

00:04:24.639 --> 00:04:31.199
accessible than ever no word yet though on whether google will take it a step

00:04:29.199 --> 00:04:34.960
further and launch google interpretive dance translator

00:04:32.960 --> 00:04:38.720
i don't know what i'm doing ignore me

00:04:36.400 --> 00:04:42.000
are you racing against the clock as a freelancer trying to start your

00:04:40.320 --> 00:04:45.040
challenging but rewarding interpretive dance company with the growth of the

00:04:43.759 --> 00:04:49.120
internet there's never been more opportunities for these self-employed to

00:04:47.280 --> 00:04:52.400
meet this need check out freshbooks cloud accounting software designed for

00:04:50.960 --> 00:04:56.479
the way that you work it's the simplest and easiest way

00:04:54.400 --> 00:05:00.720
to be more productive organized and more importantly get paid quickly

00:04:59.360 --> 00:05:05.120
you can create and send professional looking invoices in less than 30 seconds

00:05:03.199 --> 00:05:08.639
which is super important and you can set up online payments with

00:05:06.800 --> 00:05:13.360
just a couple clicks and get paid up to four days faster you can even see when a

00:05:11.280 --> 00:05:17.759
client has seen your invoice so there's no more guessing games freshbooks is

00:05:15.440 --> 00:05:22.320
offering a 30-day unrestricted free trial to our viewers to claim it go to

00:05:20.759 --> 00:05:25.520
freshbooks.comtechwiki and enter techquickie in the how did you hear

00:05:23.759 --> 00:05:30.320
about us section thanks for watching this video don't forget to like it or

00:05:27.600 --> 00:05:33.759
dislike it uh get subscribed check out our other channels

00:05:31.680 --> 00:05:37.919
and don't forget that i'm the worst dancer that has

00:05:38.560 --> 00:05:44.759
wow dennis just roasted the crap out of

00:05:41.759 --> 00:05:44.759
me