qalid7 / compath Goto Github PK

Thank you for your excellent work!
I tried to use your pipeline on my HE slides. I did this on tiling stage, however, when I get through tissue segmentation stage, it seems like no checkpoint files were provided in your code. Should I first run Train_Network_Main.py and get the model weights first? And what is the difference between Generate_output_Davros.py and Generate_output.py?

Where can I find the remaining LATTICe-A annotated samples? You reported 26,960 cells annotated, but I could only find around 4,000 cells under /latticea_test_data

I'm having the following error when I try to run this:

/compath-master/cell_detect$ python Generate_Output_Main_Davros.py --data_dir=./data --sub_dir_name=./results

where ./data contains a .jpg image and parameters-detection.txt and ./results is an empty folder.

Traceback (most recent call last):
File "Generate_Output_Main_Davros.py", line 78, in
post_process=True)
File "D:\Bureau\compath-master\compath-master\cell_detect\sccnn_detection.py", line 67, in generate_output_sub_dir
post_process=post_process)
File "D:\Bureau\compath-master\compath-master\cell_detect\subpackages\generate_output.py", line 156, in generate_output_sub_dir
assert ckpt, "No Checkpoint file found"
AssertionError: No Checkpoint file found

Can you provide two command line examples for detection and classification using you code?

Thank you

Great work! Thank you for sharing this script repo publically. We would like to re-implement your networks for a research project. I have read this paper and the paper for SC-CNN and Micro-Net. However, as I'm a newbie in python, it is difficult to build up the full training pipeline according to this repo and the paper. Could you kindly share the codes for training codes for cell_class and cell_detect? A full pipline for training and inference would be preferred.