Skip to main content
. 2021 Aug 30;7(9):169. doi: 10.3390/jimaging7090169

Figure 22.

Figure 22

Examples of meshes that were reconstructed using three faces from the MICC Florence dataset. The top row shows the input images for the networks. PRN uses a single front-facing image as input, whereas our network is fed both the front and side images. The transfer trained network output is shown in the second row, the ground truth facial mesh in the third row and the PRN output in the fourth row.