search

vqdang / hover_net

Simultaneous Nuclear Instance Segmentation and Classification in H&E Histology Images.
MIT License
512 stars 219 forks source link

lack of nuclei number after inference #274

Open akidway opened 7 months ago

akidway commented 7 months ago

Hi, I'm using run_infer.py to make inference on tiles. After run script run_infer.py, I got output json and mat file. But I found that the number of nuclei is not continues in json and mat file. There's not 315 after 314. It seems like fig below. json mat

Is that a normal phenomenon? I wonder if anyone could help me. Thanks a lot.

I use command as below: python3 /home/lzh/colon/hovernet/hover_net-master/run_infer.py --gpu=0 --nr_types=6 \ --type_info_path=/home/lzh/colon/hovernet/hover_net-master/type_info.json --batch_size=16 --model_mode=fast \ --model_path=/home/lzh/colon/hovernet/pretrained-model/hovernet_fast_pannuke_type_tf2pytorch.tar --nr_inference_workers=8 \ --nr_post_proc_workers=8 tile --input_dir=/home/lzh/colon/colon-tiles/id --output_dir=/work/lzh/colon/colon-inference/id \ --draw_dot

Junjunsss commented 3 months ago

请问你解决这个问题了吗

akidway commented 3 months ago

请问你解决这个问题了吗

没有，我目前已经不用hovernet了