search

tucan9389 / PoseEstimation-CoreML

The example project of inferencing Pose Estimation using Core ML
https://github.com/motlabs/awesome-ml-demos-with-ios
MIT License
680 stars 136 forks source link

AVLayerVideoGravity.resizeAspectFill #14

Open CoderBt opened 5 years ago

CoderBt commented 5 years ago

if I set AVLayerVideoGravity.resizeAspectFill，then pose is wrong; what can I do ?please help me!

tucan9389 commented 5 years ago

Can you give me some screenshots? And If you tell me more detail about the "wrong", I can help you.

CoderBt commented 5 years ago

request?.imageCropAndScaleOption = .scaleFill； vsessionPreset: .iFrame960x540； previewLayer.videoGravity = AVLayerVideoGravity.resizeAspectFill； UNADJUSTEDNONRAW_thumb_22a UNADJUSTEDNONRAW_thumb_22b

if I change these property，the pose predicted will change； I wish predict keypoint under the full screen ；please help me，my English is poor，haha！

CoderBt commented 5 years ago

@tucan9389 please give me some suggests，thank you;

tucan9389 commented 5 years ago

It looks like a conversion issue. You can check the conversion calculation on DrawingJointView.swift and definition of aspectFill option. If you have more question, comment below.