BoneMarrow

Associated paper

Wang C.*, Huang S., Lee Y., Shen Y., Meng S., Gaol J.(2022) Deep Learning for Bone Marrow Cell Detection and Classification on Whole-Slide Images, Medical Image Analysis, vol 75, 102270, 1-15 (SCI IF=13.828, 2/113 COMP. SCI., INTER. APP.). If you use any materials here, please cite our publication.

Cloud Demo

AI inference process and results are shown in the demo video.

Device specifications

• CPU: Intel Xeon Gold 6148
• RAM: 512 GB
• GPU: NVIDIA TITAN RTX 24 GB * 4

Time consumption

• Data extraction time: 328 seconds
• AI Inference time: 336 seconds

In the cloud demo, the system gets a WSI file from the remote NAS, and hence the data extraction time takes more than the workstation demo.

Workstation Demo

AI inference process and results as follows with demo data 1M14.mrxs.

Device specifications

• CPU: Intel Core i9-7900X
• RAM: 128 GB
• GPU: NVIDIA GeForce GTX 1080 Ti 11 GB * 2

Time consumption

• Data extraction time: 6 seconds
• AI Inference time: 94 seconds

In the workstation demo, the WSI file is stored locally, and hence the data extraction time takes less than the cloud demo.

Setup

Requirerements

• ubuntu 18.04
• RAM >= 16 GB
• GPU Memory >= 6 GB
• GPU driver version >= 410.48
• CUDA version >= 10.0
• cuDNN version >= 7.4.2

Download

Execution file, configuration file, and models are download from the zip file. (For reviewers, please use the manuscript number M...........R1 as the password to decompress the file.)

File structure

BoneMarrow/
│
├── BoneMarrow - execution file
├── setting.json - configuration file
│
├── TestImgTemp/ - temp data extraction folder
|
├── Data/ - inference data location
│   ├── 1M05.mrxs
│   ├── 1M05/
│   │   ├── Index.dat
│   │   ├── Slidedat.ini
│   │   ├── Data0000.dat
│   │   ├── Data0001.dat
│   │   ├── Data0002.dat
│   │   ├── Data0003.dat
│   │   ├── Data0004.dat
│   │   │       ⋮
│   │   └── Data0030.dat
│   │
|   ├── 1M14.mrxs
|   └── 1M14/
│       ├── Index.dat
│       ├── Slidedat.ini
│       ├── Data0000.dat
│       ├── Data0001.dat
│       ├── Data0002.dat
│       ├── Data0003.dat
│       ├── Data0004.dat
│       │       ⋮
│       └── Data0035.dat
│
├── Model/ - contains detection models
|   ├── BMntu10mix_9_i90w
|   └── NTU_ROI_i10w
|
└── Result/ - inference result is saved here

Inference

Open the setting.json file to set up the input WSI filename and the GPUs to use.
The file format is as follows:

{
    "DATA": "1M14.mrxs",    //the input WSI filename.
    "GPU": [0, 1]           //the ID(s) of GPU(s) to use for testing.
}

Then in a terminal run:

./BoneMarrow

License

This extension to the Caffe library is released under a creative commons license, which allows for personal and research use only. For a commercial license please contact Prof Ching-Wei Wang. You can view a license summary here:
http://creativecommons.org/licenses/by-nc/4.0/

Contact

Prof. Ching-Wei Wang

cweiwang@mail.ntust.edu.tw; cwwang1979@gmail.com

National Taiwan University of Science and Technology