【SAM系列】Auto-Prompting SAM for Mobile Friendly 3D Medical Image Segmentation

灵魂跑者

 关注

阅读 9

2024-01-02

论文链接：https://arxiv.org/pdf/2308.14936.pdf
在这里插入图片描述
核心：

  1. finetune SAM,为了不依赖外部prompt，通过将深层的特征经过一个编-解码器来得到prompt embedding；
  2. finetune完之后做蒸馏

相关推荐

【SAM系列】Relax Image-Specific Prompt Requirement in SAM

三分梦_0bc3 7 0 0

【论文阅读笔记】Sam3d: Segment anything model in volumetric medical images[

拾光的Shelly 19 0 0

文献速递：文献速递：基于SAM的医学图像分割--SAM-Med3D

素锦时年_1b00 13 0 0

M3D: Advancing 3D Medical Image Analysis with Multi-Modal Large Language Models

萨科潘 33 0 0

论文阅读：Low-Shot Part Segmentation for 3D Point Clouds via Pretrained Image-Language Models

木樨点点 26 0 0

3D人体重建从单张图中-DeepHuman: 3D Human Reconstruction from a Single Image

数数扁桃 32 0 0

2020-Point attention network for semantic segmentation of 3D point clouds

想溜了的蜗牛 116 0 0

论文阅读：Few-shot 3D Point Cloud Semantic Segmentation

不会弹吉他的二郎腿 69 0 0

(D3DFR)Accurate 3D Face Reconstruction with Weakly-Supervised Learning From Single Image to Image...

zhyuzh3d 126 0 0

Lightweight Attention Module for Deep Learning on Classification and Segmentation of 3D Point Clouds

云卷云舒xj 111 0 0

精彩评论（0）

0 0 举报