greenace92 1 Posted July 17, 2015 Share Posted July 17, 2015 There are these tasks on a website, and it involves using your eyes, recognition, I guess they don't have OCR or some sort of software that can identify objects... See attached photo as an example where your job is to write what you see on a receipt. So... I was thinking... each one of these pays $0.08 and if I can automate that job, then I could make money while I was working my real job. So I would think about how I understand letters (my own brain) and then figure out how to write the algorithm (is that right?) top down for example, a block is a sentence, then a smaller block is a word, then a smaller block is a letter, then the smallest detail would be the pixels comprising the block comprising the letter.... Anyway, can I get the data of a pixel? I read in this post that you can't. Why? Who said no? http://stackoverflow.com/questions/14202476/how-can-i-get-the-color-of-any-specified-pixel-on-a-web-page Does the browser take over the display on a computer and thus defines what you see or does the computer do anything with the display so that in this case my application would run on my end or perhaps a server... Link to post Share on other sites
scootstah 104 Posted July 17, 2015 Share Posted July 17, 2015 I would probably look for an existing OCR solution. But, you can get pixel data using imagecolorat(). Link to post Share on other sites
greenace92 1 Posted July 18, 2015 Author Share Posted July 18, 2015 Thanks for your response. This was a random thought. I mean that is a php function which would could be independent on their website right? Depending on windows server vs. linux? Although when I think about it, the page has to be parsed on my end in order for me to "do something with it" eg. see it. I don't know... anyway thanks for your input. I'll look that over. Link to post Share on other sites
scootstah 104 Posted July 18, 2015 Share Posted July 18, 2015 The function is part of the PHP GD library, which needs to be installed. I don't believe it is a default package, not sure. Has nothing to do with Windows or Linux. Link to post Share on other sites
0x00 3 Posted October 29, 2015 Share Posted October 29, 2015 I use Tesseract for parsing receipts and OpenCV for other related things... Link to post Share on other sites
greenace92 1 Posted November 1, 2015 Author Share Posted November 1, 2015 Is it good? I zoomed into a letter, average scale like readable from a foot or so away from a screen. The letter is say a quarter of an inch tall, when you zoom in on that, it really becomes hard to tell what letter it is. I was going to start trying to write an "algorithm" or something that would be able to tell what letter/number something was by following a top-down pixel shade sorting thing but man... haha that's pretty deep... I keep trying to figure out what to focus on and spend time on that will actually return monetary gains. I know I should pursue something to be happy, pure passion, but at this point in life, my life is not secure so I'm trying to figure something out to help my situation. Link to post Share on other sites
Recommended Posts
Archived
This topic is now archived and is closed to further replies.