mateogianolio / ocr Goto Github PK

I have a bunch of images with text in a specific font that I would like to OCR
Is it possible to create my own training data and how would one go about to do that?

https://github.com/netpi/baidu-ocr-api

I found main.js always exit when loop to 32

for(index = 0; index < samples; index++)
  {
  	captcha.generate(settings, generate,index);
  	console.log(index);
  }

My operation system is Windows 10.

I added a test.js to load it, but I got the error:

var ocr = require('./ocr.js');

/Users/xxxx/projects/ocr/mlp-character-recognition/ocr.js:1
(function (exports, require, module, __filename, __dirname) { function (input) {
^
The generated file is an anonymous function.

function (input) {
F = {
0: 0,
1: 0,
2: 0,
3: 0,
4: 0,
5: 0,
6: 0,

Dears

I've installed the lib and when I run the main.js, I get the following error :

$ node main.js
reading config file ...
... done

generating images ...
... done

layers:
input: 256 neurons.
hidden: 40 neurons.
output: 4 neurons.
learning rate: 0.03

training with 200000 inputs ...
progress: 0 %
progress: 10 %
progress: 20 %
progress: 30 %
progress: 40 %
progress: 50 %
progress: 60 %
progress: 70 %
progress: 80 %
progress: 90 %
... done (4 min 46 s 712 ms)

undefined:11137
F[256] = F[257];,F[257] = F[258];,F[257] += F[0] * F[259];,F[257] += F[1] * F[
^
SyntaxError: Unexpected token ,
at Function (native)
at Perceptron.Network.standalone (/Users/christopherbourez/apps/mlp-character-recognition/node_modules/synaptic/src/network.js:489:12)
at /Users/christopherbourez/apps/mlp-character-recognition/main.js:195:49
at null. (/Users/christopherbourez/apps/mlp-character-recognition/node_modules/pngjs/lib/png.js:84:13)
at g (events.js:199:16)
at emit (events.js:107:17)
at null. (/Users/christopherbourez/apps/mlp-character-recognition/node_modules/pngjs/lib/png.js:52:14)
at emit (events.js:107:17)
at Parser._reverseFiltered (/Users/christopherbourez/apps/mlp-character-recognition/node_modules/pngjs/lib/parser.js:358:10)
at emit (events.js:118:17)