Recognizing these gaps, further research has led to the proposal of Type II models (Fig. 1c), such as SegVol 41, which aim to extend the SAM principles to 3D spaces by replacing 2D convolutional ...