Fig. 3: Pretraining on 40,000 DFAUST meshes improves reconstruction fidelity in single pose, multi-identity data. | npj Digital Medicine

Fig. 3: Pretraining on 40,000 DFAUST meshes improves reconstruction fidelity in single pose, multi-identity data.

From: 3D convolutional deep learning for nonlinear estimation of body composition from whole body morphology

Fig. 3

Left: Evaluation set mesh reconstruction after training 400 epochs from random initialization. Test set reconstruction error was 22.7 mm. Right: Same mesh reconstruction using a model pretrained first on 200 epochs with DFAUST only followed by 400 epochs of finetuning. Test set reconstruction error was 2.0 mm.

Back to article page