isyangshu / MambaMIL

[MICCAI 2024] Official Code for "MambaMIL: Enhancing Long Sequence Modeling with Sequence Reordering in Computational Pathology"
58 stars 5 forks source link

如何传参数获得 512*512 patch (20x)呢? #5

Closed blz822 closed 4 months ago

blz822 commented 5 months ago

尊敬的作者,

   您好!请问 CLAM 在 处理 LUAD 数据集时为了获得 512*512 patch (20x),create_patches_fp.py和extract_features_fp.py的参数应该怎么设置呢?

期待您的回复。

python create_patches_fp.py --source DATA_DIRECTORY --save_dir RESULTS_DIRECTORY --patch_size 256 --seg --patch --stitch CUDA_VISIBLE_DEVICES=0 python extract_features_fp.py --data_h5_dir DIR_TO_COORDS --data_slide_dir DATA_DIRECTORY --csv_path CSV_FILE_NAME --feat_dir FEATURES_DIRECTORY --batch_size 512 --slide_ext .svs

wyhsleep commented 5 months ago

Please set --patch_size 512 and make sure the h5 file you got contains coordinate for 512*512 patches.

blz822 commented 5 months ago

我之前一直在使用DSMIL做预处理,通常还需要设置level,在CLAM中有类似的参数 patch_level ,这个参数使用 0 就可以了吗

wyhsleep commented 5 months ago

level = 0 for TCGA slides is 40X I think, you can refer to CLAM for more details.