InputImage from "fromMediaImage" method will lead Internal error after feed into FaceDetector while "fromBitmap" work as expected

[rideflag-yankai-tian]

Platform: Android

Error occurs when feed InputImage from Camera2 by using InputImage.fromMediaImage() into FaceDetector.

com.google.mlkit.common.MlKitException: Internal error has occurred when executing ML Kit tasks at com.google.mlkit.common.sdkinternal.ModelResource.zza(com.google.mlkit:common@@16.0.0:28) at com.google.mlkit.common.sdkinternal.zzn.call(Unknown Source:6) at com.google.mlkit.common.sdkinternal.zzm.run(com.google.mlkit:common@@16.0.0:5) at com.google.mlkit.common.sdkinternal.zzq.run(com.google.mlkit:common@@16.0.0:3) at android.os.Handler.handleCallback(Handler.java:883) at android.os.Handler.dispatchMessage(Handler.java:100) at com.google.android.gms.internal.mlkit_common.zzb.dispatchMessage(com.google.mlkit:common@@16.0.0:6) at android.os.Looper.loop(Looper.java:359) at android.os.HandlerThread.run(HandlerThread.java:67) Caused by: java.lang.IllegalStateException: Image is already closed at android.media.Image.throwISEIfImageIsInvalid(Image.java:72) at android.media.ImageReader$SurfaceImage$SurfacePlane.getBuffer(ImageReader.java:965) at com.google.android.gms.vision.Frame$Builder.setPlanes(com.google.android.gms:play-services-vision-common@@19.1.0:14) at com.google.mlkit.vision.face.internal.zza.zza(com.google.android.gms:play-services-mlkit-face-detection@@16.0.0:56) at com.google.mlkit.vision.face.internal.zza.zza(com.google.android.gms:play-services-mlkit-face-detection@@16.0.0:85) at com.google.mlkit.vision.face.internal.zza.run(com.google.android.gms:play-services-mlkit-face-detection@@16.0.0:164) at com.google.mlkit.vision.common.internal.MobileVisionBase.zza(com.google.mlkit:vision-common@@16.0.0:23) at com.google.mlkit.vision.common.internal.zzb.call(Unknown Source:4) at com.google.mlkit.common.sdkinternal.ModelResource.zza(com.google.mlkit:common@@16.0.0:26) at com.google.mlkit.common.sdkinternal.zzn.call(Unknown Source:6)  at com.google.mlkit.common.sdkinternal.zzm.run(com.google.mlkit:common@@16.0.0:5)  at com.google.mlkit.common.sdkinternal.zzq.run(com.google.mlkit:common@@16.0.0:3)  at android.os.Handler.handleCallback(Handler.java:883)  at android.os.Handler.dispatchMessage(Handler.java:100)  at com.google.android.gms.internal.mlkit_common.zzb.dispatchMessage(com.google.mlkit:common@@16.0.0:6)  at android.os.Looper.loop(Looper.java:359)  at android.os.HandlerThread.run(HandlerThread.java:67) 

In contrast, If convert preview frame to bitmap then create InputImage by using fromBitmap, everything works fine.

[thatfiredev]

@rideflag-yankai-tian it would be helpful if you could provide a Minimum Complete Verifiable Example Code so that we can see how you're using fromMediaImage().

[cs-googler]

Hi, could you provide some code snippets that show where you call image.close()? It is likely that you close the image too soon before mlkit is done with processing it.

[calren]

Hey @rideflag-yankai-tian, have you gotten a chance to follow up on this yet? This seems like a problem around when image.close() is called, so we were hoping you could provide the relevant code snippets to help us verify that.

[calren]

Hey @rideflag-yankai-tian , this issue hasn't had any activity for a month, so we'll consider it resolved for now! Feel free to reopen this issue if you are still able to reproduce it

[Byte-Snipper]

Hey @calren, I had the same problem. And below is some code snippets.
Image image = reader.acquireLatestImage(); InputImage inputImage = InputImage.fromMediaImage(image, 0); image.close(); //handle inputImage here

[cs-googler]

Hi, the image is used for processing in the ML pipeline. You should close it in the detector's onComplete() listener instead to make sure the pipeline does not need it anymore.