Fig. 4 | Scientific Reports

Fig. 4

From: Osteoporosis prediction from hand X-ray images using segmentation-for-classification and self-supervised learning

Fig. 4

Two-stage training for osteoporosis prediction. In the first stage, augmented samples are processed through an encoder for self-supervised pretraining as part of a pretext task. The backbone classification encoder network (gray trapezoid with slanting lines) is subsequently fixed and repurposed to address the downstream classification task by incorporating a trainable linear layer (cyan trapezoid).

Back to article page