zhixuhao / unet

unet for image segmentation
MIT License
4.6k stars 2k forks source link

How to perform multi-class detection? #103

Open zhouhao-learning opened 5 years ago

zhouhao-learning commented 5 years ago

Hello, I am very happy to see your code. Now I have a piece of data. There are 2 types in total. One is benign data, the other is malignant data. They all have corresponding mask map labels. I want to implement multi-classification through Unet. Detecting tasks, but I don't know how to modify the model, and how to process my data for multiple classifications. I don't understand how my label should be defined. Can you teach me more? Thank you!

wuyang0329 commented 5 years ago

Hello, I am very happy to see your code. Now I have a piece of data. There are 2 types in total. One is benign data, the other is malignant data. They all have corresponding mask map labels. I want to implement multi-classification through Unet. Detecting tasks, but I don't know how to modify the model, and how to process my data for multiple classifications. I don't understand how my label should be defined. Can you teach me more? Thank you!

I think my project would help you: https://github.com/wuyang0329/unet

Sam813 commented 5 years ago

@zhouhao-learning Have you done the multi-class segmentation? I have the same problem here

harsh52 commented 5 years ago

Hello, I am very happy to see your code. Now I have a piece of data. There are 2 types in total. One is benign data, the other is malignant data. They all have corresponding mask map labels. I want to implement multi-classification through Unet. Detecting tasks, but I don't know how to modify the model, and how to process my data for multiple classifications. I don't understand how my label should be defined. Can you teach me more? Thank you!

did you find the solution? i have the same issue. please ping me if find anything.

wiqarahmad789 commented 5 years ago

Hello respected i have my data of weather maps with multiple kinds of line and i need to detect just dashed line through unet ..kindly guide a little i will be very thanklfull.