A lightweight SAM student trained with MSE plus perceptual loss reaches SAM-like Dice scores on some medical datasets but falls behind on breast ultrasound, with no error bars or prompt details.
This approach addresses the computational limitations of SAM’s Vision Transformer (ViT) encoder by distilling its knowledge into a lightweight ResNet [12] based encoder
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
method 1
citation-polarity summary
fields
eess.IV 1years
2025 1verdicts
REJECT 1roles
method 1polarities
use method 1representative citing papers
citing papers explorer
-
Efficient Knowledge Distillation of SAM for Medical Image Segmentation
A lightweight SAM student trained with MSE plus perceptual loss reaches SAM-like Dice scores on some medical datasets but falls behind on breast ultrasound, with no error bars or prompt details.