search

dengxl0520 / MemSAM

[CVPR 2024 Oral] MemSAM: Taming Segment Anything Model for Echocardiography Video Segmentation.
MIT License
130 stars 12 forks source link

dataset and program #6

Closed gloryhoper closed 5 months ago

gloryhoper commented 5 months ago

Dear author,I had the following problems when running the program:

  1. Failed to run from utils.tools import corr, bias, std, without corresponding functions or variables in utils.tools. 2.FileNotFoundError: [Errno 2] No such file or directory: '/tmp/CAMUS_public_256/class.json' After running preprocess_camus.py, the class.json file is not found in the generated folder and the original file, and I want to know how to get this file. Finally about the data set. For now, we can only use CAMUS and EchoNet-Dynamic. For how to use our own avi format video dataset or jpg format image dataset, I wonder if the author will update it?
Zong-Liang commented 5 months ago

MARK! SAME PROBLEM!

dengxl0520 commented 5 months ago

I'm sorry for the problems you're having, here are some answers to your questions:

  1. Regarding the missing corr bias std function in utils.tools, you can add the corresponding function by yourself or comment out the corresponding evaluation code temporarily. I will update the utils.tools later.
  2. The file class.json comes from SAMUS and its contents should be: 
    {
"EchoNet": 2,
"CAMUS": 2
}
  3. Due to time constraints, I may not update this feature.
gloryhoper commented 5 months ago

Thank you for your reply.I understand.