I want to find images from the database using a webcam.
Specifically, I want to set a "price kiosk" where people can walk with one item, put it in front of the camera, and search for the database on the price. Cause reasons (most easily used in use) I do not want to use barcodes in products.
Items are relatively easy to scan (they, for practical purposes, 2D: they are comic books). I have all the scans already scanned So what I like is a way to take an image from the webcam and use it as a source of search. Of course the image will get distorted (angle, focus, resolution, light, rotation, etc.) Google Goggles There is no problem for), because I scanned the comic book has been included in many situations and it is capable of finding them.
Now, I am doing some research, I have done a very good job with OpenCVI, so I think it should not be too difficult to implement. Particularly considering my datasets is much smaller than Google (approximately 2000 different products).
What specifically am I seeing? Object detection, recognition, attributes ...? I'm confused and I do not even know where to start from.
read on SIFT: (Scale Irrevocable Facility Transform)
No comments:
Post a Comment