Using Wikitude SDK 9.6 Unity Professional Edition with Unity 2019.4.20 LTS version.
I have multiple random image url which can be added at runtime. Currently with wikitude image tracker I can add one url as target collection which is the problem.
I don't have any pre-made zip or wtc file. Also using multiple image tracker is not an option as I don't know which image will be in front of the camera and which image tracker to enable.
Is there any way I can have simultaneous image recognition in this case for multiple random image url given at runtime?
is there a limitation why you can't generate zip or wtc files?
Creating a zip archive during runtime with the images could be an option.
I will bring this up for discussion internally, if it is feasible to support multiple image urls and texture assets as well...
First, thank you for your response and considering my need.
Second, in my system, target image URLs can be set even in post release which can be any JPG/PNG image URL and set with a particular prefab content. So I don't know which images will be chosen.
Third (thank you for suggesting), I have considered creating zip archive during runtime. That would be my last resort if I don't find any better solution.
For now, in my case, the best possibility from Wikitude I see if target collection can take a list of URLs. Since it can take an image from a URL and track it, the same should be true for multiple images if target collection can take it.
With the asset bundle workflow, .zip/.wtc files are also working fine (as a separate upload). You anyhow need to link an image URL to a prefab right? That can also be done for a target name in a .zip/.wtc target collection right?
I think using .zip/.wtc files is still the safest option here.
Thank you again for your suggestion. Actually linking with a prefab is the easy part. I am using Unity's Addressable asset address and image address (currently direct image URL) and saving in a JSON file.
The hard part I am finding, is the combination of images which will be chosen as target in front of the camera. It can be any combination of images. If I go with premade .zip/.wtc, either I have to have a giant collection of images where all target images will be inside (I donno how many there will be) or create multiple .zip/.wtc files according to the set of images.
For the first option i mentioned, can be risky as I donno how many images there will be.
For the second, there is a possibility of multiple images being in multiple .zip/.wtc files. In which case, I may have to use multiple Image Tracker (with single .zip URL for each) and let the user change manually from image sets which has the possibility to make the user confused.
So, that's why in my situation, I would prefer simultaneous detection from multiple URL (even multiple .zip file is fine). Before you mentioned, about texture support that would be also awesome if Unity's native Texture2D can be supported. If there is any possibility of either of the options please let me know.
creating a .wtc file with all images might not be that big at all. Another option would also be to use our Cloud Recognition service. Is that an option?
Haven't considered nor tested with Cloud Recognition Service yet. Will give it a shot. Thank you.