Is it possible to track face movement from local video using ARKit?
Since ARKit face tracking requires some 3d information, I guess it might very hard.
However, to send ARKit-optimized tracking data to unreal, looking for local adaptation of ARKit Face tracking.
Related
I am using an app that was developed using the ARKit framework. More specifically, I am interested in the 3D facial mesh and the face orientation and position with respect to the phone's front camera.
Having said that I record video with subjects performing in front of the front camera. During these recordings, I have noticed that some videos resulted in inaccurate transformations with the face being placed in the back of the camera whereas the rotation being skewed (not orthogonal basis).
I do not have a deep understanding of how the TrueDepth camera combines all its sensors to track and reconstruct the 3D facial structure and so I do not know what could potentially cause this issue. Although I have experimented with different setups e.g. different subjects, with and without a mirror, screen on and off, etc. I still have not been able to identify the source of the inaccurate transformation. Could it be a combination of the camera angle interfering with the mirror?
Below I attached two recordings of myself that resulted in incorrect (above) and correct (below) estimated transformations.
Do you have any idea of what might be the problem? Thank you in advance.
I have faced the issue of real face detection using Vision Framework.
I have referred below apple link.
https://developer.apple.com/documentation/vision/tracking_the_user_s_face_in_real_time
I used demo code provided in above link. I see, Camera can detect the face from printed photo or passport photo. It is not real face photo. How can I know if this is not real face in camera using Vision framework?
You can use https://developer.apple.com/documentation/arkit/arfacegeometry
This will create a 3D mesh of a human face. A 3D mesh will have different values (e.g. vertices , triangleIndices), in its topology compared to a 2D picture.
Here is a project link
here I have used camera API for face detection and eye blinking. you can check and customize according to your requirement.
Update: Here is another project for liveness Check using MLKit link
Vision + RealityKit
Apple Vision framework has been processing "2D requests". It works only with RGB channels. If you need to process 3D surfaces you have to implement LiDAR scanner API, that based on Depth principles. It will allow you to distinguish between a photo and a real face. I think that Vision + RealityKit is the best choice for you, because you can detect a face (2D or 3D) at first stage in Vision, and then using LiDAR, it's quite easy to find out whether normals of polygonal faces are directed in the same direction (2D surface), or in different directions (3D head).
I like to know which technology I can use for creating android application which has the feature of detecting an random object like hoarding board and after detecting, a video must played over that object. If the user move the camera , the video must also move over that object position.
I have an example of what I want to achieve in this video of big burger https://www.youtube.com/watch?v=lhXW8_7CaHM . But I want to play that video on any type of hoarding the camera detect. Not only on specific trained object.
I have study some technology like tensorflow and Vuforia which can made ar application.
But I'm not sure if it can detect real time objects with tracking of the objects position.
I am trying to wrap my head around Vuforia's capabilities. I want to make an app which lets me place a 3D object into a camera view and have that 3D object stick to the world. I've been learning how to use Vuforia in Unity3D, and Vuforia seems to be slightly capable of this, but is severely limited by its craving for "Targets". It doesn't seem to be able to do much if I don't give it some sort of target.
One workaround I've found is to set the ARCamera's World Center Mode to DEVICE_TRACKING. This seems to let me place a 3D object into the world (in Unity) and have this object overlay into the camera feed, almost making it seem like it's anhcored to the real world. This doesn't work perfectly though: it tracks properly when I angle the device up/down/left/right (rotation), but it does not seem to track the device's translational motion; that is, when I move the device forward/back/left/right, the overlaid object doesn't get closer/farther nor does it rotate as I move around it.
Is it possible to get this sort of tracking out of Vuforia, or am I better off switching to something like Google Tango?
The difficulty with setting World Center Mode to CAMERA in Vuforia is that apparently 3D objects rotate around the camera based on its accelerometer/gyroscope changes. This doesn't allow for objects to be anchored to the environment. Instead they follow with the camera.
Kudan is a good markerless tracking option.
I am going to build a FPS video game. When I developing my game, I got this problem in my mind. Each and every video game developer spend very big time and use a lot of effort to make their game's environment more realistic and life-like. So my question is,
Can We Use HD or 4K Real Videos as Our Game's Environment? (As we seen on Google Streetview - but with more quality)
If we can, How we program the game engine?
Thank you very much..!
The simple answer to this is NO.
Of-course, you can extract texture from the video by capturing frames from it but that's it. Once you capture the texture, you still need a way to make a 3D Model/Mesh you can apply the texture to.
Now, there have been many companies working on video to 3D model converter. That technology exist but is more for movie stuff. Even with this technology, the generated 3D models from a video are not accurate and they are not meant to be used in a game because they end up generating a 3D model with many polygons, that will easily choke your Game engine.
Also, doing this in real-time is another story. So you will need to continuously read a frame from the video, extract a texture from the video, generate a mesh with the HQ texture, cleanup/reduce/reconstruct the mesh so that your game engine won't crash or drop many frames. You then have to generate a UV for the mesh so that the extracted image can be applied to the current mesh.
Finally, each one of these are CPU intensive. Doing them all in series,in real-time, will likely make your game unplayable.I also made doing this sound easy but it's not. What you can do with the video is to use it as a reference to model your 3D environment in a 3D application. That's it.