0
$\begingroup$

I want to implement a real-time system for image comparison (e.g. compare a face with a reference one) on an Odroid. I would like to know what are the most suitable architectures for this task. I started with methods based on triplet loss (like Facenet) but I realized that a real-time solution is not feasible. Are there good, light alternatives?

$\endgroup$
1
$\begingroup$

The problem might not be caused by your loss function. Deep learning models tend to be computationally demanding. Mobile devices are not usually prepared for handling models with high throughput. That being said, you might try to:

  1. Prepare smaller model - less layers, less computation

  2. Mobile models optimization - Google provides some materials on optimizing Tensorflow models for mobile inference:

    https://www.tensorflow.org/mobile/prepare_models https://www.tensorflow.org/mobile/optimizing

| improve this answer | |
$\endgroup$
0
$\begingroup$

For face ID, Apple is using Siamese Network. You may get a better idea here https://towardsdatascience.com/one-shot-learning-face-recognition-using-siamese-neural-network-a13dcf739e

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.