Which methods are used in Android cameras for face detection?

I want to know how can I take an image from android process it by opencv and then return the result to android.

How do we calculate misclassification in face recognition system?

Am working on face recognition in video, i want to start work on deep learning based face recognition for better accuracy.

kindly suggest me if i use deep learning based method will it improve the accuracy?

currently am using modified version LBP, ELBP, SIFT, Local Directional based feature extraction.

I have trained my own dataset and I have got frozen inference graph model ,

I have converted this model to tflite model to use it with Android SDK.

I need to know how can I deploy tflite model and how can I extract predictions from tflite model?

I've seen accuracy [(2-FAR-FRR)*100/2] reaching to about 80% to 85%. Could anyone please share with me the greatest accuracy you know in off-line HSV and also specify the reference from which the value was obtained?

Also do any of these methods address the individual variability of signers? Eg., an expert signer would have a greater self similarity in his/her signature, while a novice signer would have a low self similarity in signing his/her own genuine signature. Can any of those highly accurate HSV systems retain their accuracy in both these cases? If so, how? Are there any other limitation to such systems?

I do believe the contribution to these questions would benefit many in this field of study.

Human detection is mostly done via image processing of video/camera data. I'm searching for simpler and more direct solutions e.g. passive infrared or electromagnetic fields.

The methods for face detection that work well in unsuitable lighting condition.