- ABOUT US
- LATEST NEWS
- CONTACT US
The AI automatically calculates a physical weight (gram) of pneumonia, as well as the extent (%), which is in accord with the use case of ACR(American College of Radiology)'s Data Science Institute. Users can segment and analyze COVID19 pneumonia on CT with simple commands.
This software works on Windows7 64bit or higher operating system and NVIDIA graphics card(with 2GB GPU memory or higher) supporting CUDA and latest drivers.
Please make sure of the update of the GPU driver with the latest version before use.
Environment Detail : OS-Microsoft Windows 7(64bit) or higher, CPU-Intel i5 or higher, RAM-8GB or higher, GPU-NVIDIA GeForce 1000series or higher(with 2GB GPU memory or higher)
Please contact us (COVID19@medicalip.com) if you have any problems or do not have any computing resources.
- Improved reinforcement learning model
- Improved report template
- Bugfix on interface and analysis
- Improved report template
- Improved installer and license dialog
- Improved 3D opacity visualization
- Bugfix on report tab update
- GPU hardware inspection exception handling
- Improved 2D visualization
- DICOM(CT) loading
- Automatic lung, pneumonia segmentation by deep neural network
- Automatic feature extraction: volume, mean HU, standard deviation, pneumonia burden, % extent
- Automatic report generation
The manuscript dealing with the technique for this AI application will be submitted to the journal this month. The software currently can be used for research purposes only. We are waiting for collaborative researchers to verify the value of whole-lesion quantification on CT in COVID19 (contact: COVID19@medicalip.com).
The Spearman correlation coefficients for % extent and pneumonia burden between AI and human expert were 0.970 and 0.982, respectively, in the internal validation dataset.
v188.8.131.52 : 0.970, 0.982
v184.108.40.206 : 0.962, 0.924
The manuscript dealing with the technique for this AI application will be submitted to the journal this month. Please cite the MEDIP as follows: MEDIP COVID19 v220.127.116.11 (MEDICALIP, Co. Ltd. Seoul, Korea)
1. Extension of Coronavirus Disease 2019 (COVID-19) on Chest CT and Implications for Chest Radiograph Interpretation
Hyewon Choi / Xiaolong Qi / Soon Ho Yoon / Sang Joon Park / Kyung Hee Lee / Jin Yong Kim / Young Kyung Lee / Hongseok Ko / Ki Hwan Kim / Chang Min Park / Yun-Hyeon Kim / Junqiang Lei / Jung Hee Hong / Hyungjin Kim, MD / Eui Jin Hwang / Seung Jin Yoo / Ju Gang Nam / Chang Hyun Lee / Jin Mo Goo
Radiology: Cardiothoracic ImagingVol. 2, No. 2, Mar 30 2020
This project was initiated by the proposal of the CHESS-Korea Deep Neural Networks of CT Imaging for COVID19 with the support of MEDICALIP.
List of current contributors is as follows and more contributors will be added:
Xiaolong Qi, The First Hospital of Lanzhou University, China and Soon Ho Yoon, Seoul National University Hospital, Korea
Seung-Jin Yoo, Hanyang University Medical Center, Korea
Shohei Inui, Japan Self-Defense Forces Central Hospital, Japan
Yeon Joo Jeong, Pusan National University Hospital, Korea
Kyung Hee Lee, Seoul National University Bundang Hospital, Korea
Young Kyung Lee, Seoul Medical Center, Korea
Bae Young Lee, Eunpyeong St. Mary's Hospital, Korea
Jin Yong Kim, Incheon Medical Center, Korea
Kwang Nam Jin, Seoul Metropolitan Government-Seoul National University Boramae Medical Center, Korea
Jae-Kwang Lim, Kyungpook National University Hospital, Korea
Yun-Hyeon Kim, Chonnam National University Hospital, Korea
Ki Beom Kim, Daegu Fatima Hospital, Korea
Zicheng Jiang, Ankang Central Hospital, China
Chuxiao Shao, Lishui Central Hospital, China
Our next step is to develop free software for quantifying COVID19 on chest X-ray radiographs in two months.
If you agree with our initiative, please join our step as a co-researcher.
Any resource support or collaboration is welcome to speed up this global pandemic troubleshooting.