534634 Image Recognition with Augmented Reality
N/A
Paid on delivery
The project consists of two components 1. a user app and 2. a server side application.
The app (intially iOS with a later requirement for Android) will use the camera within the smartphone to view printed pages and recognise pre determined images, once recognised the app will float buttons over the images from the live camera in an augmented reality manner.
Each recognised image will require 2 buttons. One button when selected will launch an internal browser within the app to call up additional information from a web server. The other button will add the data from the recognised image to a “basket”.
The browser will require a “done” button to return to the camera view. There will also need to be a button to show the contents of the basket.
There may be more than one image on screen at one time requiring buttons to float over each image.
A few areas that need defining will be how the images are recognised, as I would prefer this not to be a callout to an external database but rather by referencing a pattern from the image that could be syncronised when the app is launched. The issue here is that there could be over 500 images that require recognising. This would be refreshed once a week. Therefore the reference data must be small.
The concept is to be able to recognise images instantly, or quicker than a callout and database search. The image recognition will be like for like, that is the exact image on the printed page will be the one referenced. The other side of this is the synchronisation should take around 30 seconds maximum
The server side requires the ability to call up a PDF and then the operator would select the image area that is required to be referenced. The selection process could be by using a third party application such as Adobes Illustrator or Photoshop. Once selected the images are tagged with a unique reference code and placed on the server database, these images are to be used as the reference for the recognition from the app.
The requirement of this project is to create a solution that on the app side recognises images via the camera “viewing” a printed page then referencing a database almost instantaneously creating an augmented reality experience with buttons “floating” over each image that is recognised (there could be multiple images in view, requiring multiple buttons) The buttons would then invoke either a “basket” to record the image selected and another button that makes a callout via an internal browser (this is also important as there is a need to go back and forward from the view on the camera and the browser)
The images will be prepared ahead of time delivered to the server as the images to be recognised. This may involve 30 pages with up to 15 images per page. The manual aspect of this is not an issue, however the tagging and in-gestation into the server should be simple and quick.
Project ID: #2280577
