Bill Upload for Comparison website?

ches32

Regular Member
Joined
Oct 30, 2014
Messages
454
Reaction score
150
Hi all

I'm in the process of rebuilding my comparison website and thought of a great function to add.

Instead of the user filling out the form with their name, address, current provider and plan, and current usage, I would add bill upload function, then connect it to Google Vision Text Detection (they have both from image and PDF) So that user don't need to fill out manually. So basically it scans image/pdf and fills out filed in the form on the website.

I tested G Vision with few bills and it seems to work great.

My question is the technical implementation, How hard is it to code something like this and what failure % is to be expected?

Have a good weekend!
 
I assume they have an API, so you would need upload function built in to your website - level of difficulty probably lies in how your site has been coded and whether its easy to integrate something new into it. Then you would need to interface with the API to get the details across to you. I imagine security would be a fairly big concern here, so you'd have to consider the implications of storing and transferring these documents.

Sounds like a great idea though.

If you want a good idea it might be best posting this in hire a freelancer - then you should get some quotes which gives you an idea of what it will cost you.

Let me know and I can move it there, unless you are planning to do it yourself and just looking for advice?
 
I assume they have an API, so you would need upload function built in to your website - level of difficulty probably lies in how your site has been coded and whether its easy to integrate something new into it. Then you would need to interface with the API to get the details across to you. I imagine security would be a fairly big concern here, so you'd have to consider the implications of storing and transferring these documents.

Sounds like a great idea though.

If you want a good idea it might be best posting this in hire a freelancer - then you should get some quotes which gives you an idea of what it will cost you.

Let me know and I can move it there, unless you are planning to do it yourself and just looking for advice?
Google vision has API and everything.

I'm in the process of building a new site, development hasn't begun yet so timing is good.

Yea I'm just looking for advice on the topic since no comparison sites in my country has this, and I only found few that has this function in US, UK or AU, some of them are manual as well, so maybe it's a lot more complicated than what it seems like.
 
Thread moved in that case.

I’d say it’s fsirly complicated - in that it needs custom development from the ground up. But I’m sure it’s doable.

Like you mentioned, errors would be a problem you’d have to deal with, but you could have the form filled and allow the user to see and check it.

I’ve seen similar facilities offered by other sites for things like grabbing details from credit and debit card via picture / camera access.

Also, I’ve moved this to programming > web design.
 
So basically scanning pdf and then applying image manipulation to extact text. Thats weak for currently in java script but its not impossible. But one of very difficult projects. I would say. As javascript image manipulation libraries are pretty weak, i would prefer that you use some kind of service. Where MATLAB appliaction is basically getting text out of image/pdf and then saving it in text file.

So it would be costly to build. BTW I am full stack web develoepr and excellent profile on good freelancing platforms. Thats my 2cents.
 
So basically scanning pdf and then applying image manipulation to extact text. Thats weak for currently in java script but its not impossible. But one of very difficult projects. I would say. As javascript image manipulation libraries are pretty weak, i would prefer that you use some kind of service. Where MATLAB appliaction is basically getting text out of image/pdf and then saving it in text file.

So it would be costly to build. BTW I am full stack web develoepr and excellent profile on good freelancing platforms. Thats my 2cents.
The extraction will be done by Google Vision (https://cloud.google.com/vision/) then I get the text in json and just insert it back in the field so user can check that extraction went smooth.

Question is how hard is it to connect all that and what failure % can be expected.
 
Back
Top