Use UNet for image segmentation, as shown in the figure above, which can be applied to medical image segmentation. This example is Chest CT image segmentation.
[Operation steps and instructions]
Before using UNet for training or inference, please make sure that the dataset folder is selected correctly.
The "browse" button next to the "Select Dataset" field can open the data folder location for the user to confirm and modify.
1.annotation labelme json: Open the annotation webpage for UNet image annotation.
2. labelme json to dataset: Convert label files to generate training files.
The converted files will be placed in the dataset folder and the gt folder.
3. delete log: delete logs folder.
After pressing 3. delete log, the logs folder in the select dataset will be deleted.
Before training, make sure that there are these four folders in the image/train folder of the selected dataset, and there are files in them, then press 4. train to start training.
dataset: Generated by running 2. labelme json to dataset.
gt: Generated by running 2. labelme json to dataset.
img: Training image.
label: A label file is generated after 1.annotation labelme json is marked.
There are several ways of inference:
5.inference (GPU): Infer a single image.
6.inference folder (GPU): Infer the selected image folder.
7.inference api (GPU): Use web pages for image inference. After pressing the "7.inference api (GPU)" button, press the "8.inference api browser" button to open the web page.
Infer a single image.
Infer the selected image folder.
The location of the result in the inference folder.
Use web pages for image inference.