Is there any documentation and/or sample code of scene recognition and sharing available? I saw the blog mention of it, but don't see any mention in the docs or samples (although I may have missed it?)
Hi ! Very excited about this new version of Wikitude! I have the same question as Patrick.
Also, on your blog post, you mention that with scene recognition, we will be able "to localizes the user in a highly precise manner. Much more than GPS does." Do you have any information to share about that ?
If you download the beta package from here https://www.wikitude.com/download/?link=beta then, you will find a documentation section inside the package. Scene Recognition is actually our improved Object Recognition feature. In SDK 8, we have improved this feature so instead of uploading a video with all the limitations this may have (video size and quality), you can upload images of the object. In addition, if you wish to upload more images at a later stage, you do now have to create a new .wto file. You simply add the images in your existing file and recreate the initial .wto.
Moreover, with Object Recognition you can now try and track bigger objects. Most of this information is included in the Object Recognition section of the beta documentation. However, we will include more information once we have the final version of the SDK 8.
What we mean here is that the localization is visual, so it will match exactly what you tracked before. Basically, this is the feature 'Persistent Instant targets' that we have included with this new version. This information is included in the beta documentation as well.
We do realize that there is still much information missing from the documentation section that we have included in the beta package. As this was a beta release we have not finalized the documentation yet but we will certainly do so once the final release is launched.
Thank you for your feedback and if you have more questions we are here to help ;)