Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add support Android 9 or higher #59

Open
Githubmaster9089 opened this issue Mar 30, 2021 · 11 comments
Open

Add support Android 9 or higher #59

Githubmaster9089 opened this issue Mar 30, 2021 · 11 comments

Comments

@Githubmaster9089
Copy link

Dear fabio914.
I am wondering if u are also developing an android version of that application.

Regards,
Githubmaster9089

@ilker-aktuna
Copy link

yes. we all wonder that.
and I am ready to pay for it :)

@Githubmaster9089
Copy link
Author

I am ready to pay money for sth. like that too

@ilker-aktuna
Copy link

actually I am able to develop android apps but I know nothing about image processing. If I had a clue on where to start I would be interested in developing this for Android.

@fabio914
Copy link
Owner

@ilker-aktuna You might want to start by taking a look at the Oculus MRC plugin for OBS and checking how the Mixed Reality video is formatted and how it's sent over the network. Then you'll want to check how to integrate FFMPEG into your Android app and how to decode the video and present it in a View. After that you can start checking how to obtain the video from the camera, and how you can compose the Mixed Reality scene (by placing the video from the camera in between the foreground and background videos). Unfortunately, there's no simple option for the "person segmentation" on Android. You can try to adapt an ML model that's typically used in a desktop application (example), but unless it's done properly and the phone has specific hardware that can run this model faster, the results will likely be very slow. You can instead play around with some chroma key shaders, to replace the color green with transparency, and apply those on top of the image from the camera, but then the user will have to use a green-screen.

@zhubinsheng
Copy link

https://github.com/ZHKKKe/MODNet
run slow ,only 2or3 fps on android

@ilker-aktuna
Copy link

@zhubinsheng
what is this ?
seems not even related to the topic.
what are you trying to tell ?

@zhubinsheng
Copy link

@zhubinsheng 这是什么? 看起来甚至与主题无关。 你想告诉什么?

I want to build an android app that is similar to your idea. Using AI model to run is too slow

@ilker-aktuna
Copy link

ok. I also "want" to build. But I don't have the courage to start from zero. So I am unable to build right now.
But what's the reason to share something not really very related to the topic ?
If there is any relation, can you explain ?

@PaulMndn
Copy link

PaulMndn commented Oct 5, 2023

@fabio914 It looks like Google has released mlkit which includes a "person segmenter", they call it "Selfie segmentation", but as I understand it, it works with any image or video, including live video. With that it should be possible to create a port for android, right? They say the API has a varying latency depending on the image resolution, but I imagine that would be quite stable for this application, and the app could just include a slider to adjust the offset of the two video sources.

Would it be possible to port this app to android using this feature? (also, possibly @ilker-aktuna)

@happyplacetraveler
Copy link

happyplacetraveler commented Oct 6, 2023 via email

@fabio914
Copy link
Owner

@PaulMndn something like might work in an Android app for Mixed Reality Capture. The issue today is that Meta stopped supporting Mixed Reality Capture (the official Mixed Reality Capture Calibration app is broken now that they've changed the Android version that they're using), and many games no longer support Mixed Reality Capture (for example, I tested Beat Saber and Synth Riders recently, and they also stopped working).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants