inspirit / jsfeat

JavaScript Computer Vision library.
MIT License
2.74k stars 372 forks source link

coin recognition #39

Closed dwinuryani closed 9 years ago

dwinuryani commented 9 years ago

my name is dwi nuryani rahmayanti and want to know about JSfeat more. actually i have a small project from my lecturer about implementing coin recognition using web application. can i use some library from JSFeat and implemented it in my case ? if yes, what library function that i should use in order to do that ? thanks for your attention before.

outer-scope commented 9 years ago

Greetings Dwi,

first of all, out of curiosity, are you from Indonesia? :)..

Anyways, i am not actually an expert on this matter, let alone a JSFeat developer. I am also currently using JSFeat for my thesis project. This library is the best I can find so far. (Thanks Inspirit!)

Anyhow, I understand the challenge on realizing the samples shown in the demos onto your work. But worry not, I have been there as well.

Judging from your question, I assume that you miss the connection from JSFeat examples and your project.

Let me give you a bit of a guidance. Since your task is coin recognition, we need to narrow our discussion down to object recognition. Here, you can already see some examples explained at http://inspirit.github.io/jsfeat/#features2d. From there, please take a look specifically at http://inspirit.github.io/jsfeat/sample_orb.html. This is your starting point! Please clone JSFeat source, if you haven't already, and you can start examining the codes.

The idea behind ORB descriptor example, is that you can "train" an image recognition. This, of course, requires you to be connected to a camera and a browser that supports getUserMedia. The ORB descriptor extracts an image and use them as a "training model". The video stream, will then be compared with the training model. If the trained object "appears" in the video, the recognition will mark their presence.

Now the tricky part is to train your model. As you can already realize, you need 2 train models: the front facing and the rear facing surface of the coin. How to implement this? You might need to spend some time to modify the codes. Have fun! :).. (Good thing is that your model is a coin..)

Last but not least, as a tip, maybe next time you might want to consider researching before posting questions. In my opinion, your question can't be categorized as an issue. It's more like asking for guidance or tutorial. Unless you have already dig in the codes and bumped to something that you cannot solve, I suggest you to avoid asking such questions.

Great! Hope this helps, if you haven't already found a solution. Please ignore anything you find unpleasant! Good luck! :)