search

JCruan519 / VM-UNet

(ARXIV24) This is the official code repository for "VM-UNet: Vision Mamba UNet for Medical Image Segmentation".
Apache License 2.0
393 stars 12 forks source link

输入通道数问题 #32

Open MQN-80 opened 3 months ago

MQN-80 commented 3 months ago

代码中输入通道数为3，但是实际都是灰度图片，这是什么原因

JCruan519 commented 3 months ago

@MQN-80 Hello, for the ISIC17 and 18 datasets, the input is RGB images. For the Synapse dataset, we follow the processing method of Swin-UNet, which you can refer to at https://github.com/JCruan519/VM-UNet/issues/30.

MQN-80 commented 3 months ago

那我可以按ISIC17格式制作自己的多分类数据集吗，改一下class就可以吗

JCruan519 commented 3 months ago

@MQN-80 Hello, the handling of multi-class tasks can refer to the approach used by Synapse.

MQN-80 commented 3 months ago

@MQN-80 Hello, the handling of multi-class tasks can refer to the approach used by Synapse. 谢谢你!但是我的是3通道多class图片，我转成灰度图后确实能跑，但是可能会影响训练效果，不知道有没有方法可以训练3通道图多类别的

mistletoe111 commented 2 months ago

我想知道config里面classes为啥设置是1 为啥不是2呢 不是两个类别吗