DeepCatch redefines your body composition in 3D. Deep learning algorithms enable you to analyze the compositions, seven layers within few minutes (3 mins for whole body, 2 mins for abdomen). Distributions of adipose tissues and muscles in 3D offer you valuable quantitative medical data for numerous application.

You can check accurate muscle mass and fat mass, also distribution from CT images. Enhanced imaging and non-enhanced imaging are available. When we design artificial intelligence, medical staff participate directly to improve its accuracy through anatomical verification.

Value Proposition

- Automated seven body compositions segmentation within 2 minutes
- DeepCatch provides significant value based on quantitative data of muscle and fat mass with 97% accuracy
- Automated L3 level and abdominal waist detection
- Provides user define report for patients and expert report for researchers


To measure TAMA and VFA, pre-contrast CT images were uploaded to DeepCatch for sarcopenia and cancer patients to find any relationship to find proper treatment.

Intended Use

Abdominal CT and Whole Body CT


In order to analyze CT image certain spec of high GPU and memory are required.

Model performance metrics

97.6% accuracy / on Contrast CT (in 3D)

Performance Curves

mPerformance Curves

Based on whole body CT analysis with contrast image, seven layers segmented in 2D and 3D.



MEDICALIP we are leading the future of healthcare with innovative technologies. We are developing various medical imaging analysis technology based on AI and medical 3D printing technology to reduce the time and cost of medical staff and the pain of patients. We are creating a future that will enable a better medical experience and service.