Face Generator

We will train a generative adversarial network (GAN) to generate new faces. This technology can be applied to data enhancement, privacy protection, movie-assisted shooting, etc.

Background

Face recognition algorithm has been developed for a long time and has been widely used in industry, but face occlusion is still an obvious problems. GAN can be used to recover occluded faces. On the other hand, pose has a great impact on face recognition and other algorithms, which will also cause the above occlusion problems. Using GAN for facial pose simulation, such as converting the side face to the front face, will greatly improve the performance of face recognition and other models.

Prepare Dataset

First step: Specify the application scenario. In this example, the focus areas of different organs in the image are divided.
Second step: It is recommended to prepare at least 200 face images. For example:

PS: Click here to download a sample dataset.

If you need a complete dataset, please send mail to hpchelp@lenovo.com

Third step: Create a dataset through LiCO.

Click [Create Dataset].

Train Model

When your dataset preparation is complete, you can click [Create Model] to complete the model creation.

And then click [Train Model] to start training.

You can also predict multiple images by [Batch Predict].