Intelligent Diagnosis for Lung Disease
- Comprehensive screening of five lung abnormalities and 120 diseases
- 96.2%
- Consolidation
- 92.7%
- Cystic shadow
- 91.9%
- Reticular shadow
- 90.5%
- Mixed-density shadow
- 95.8%
- Nodule shadow
Intelligent Diagnosis for Chest X-Ray
- Aided screening for chest X-ray, covering more than 20 symptoms of major cardiothoracic diseases
- Nodule, lump, pneumonia, bronchitis, pulmonary edema, calcified foci, atelectasis, emphysema, fibrosis, cavitation, tuberculosis, pleural effusion, pneumothorax, pleural hypertrophy, pleural calcification, tortuous aorta, aortal calcification, enlarged heart shadow, prominent pulmonary artery segment, rib fracture, destruction of rib, clavicle fracture, scapula fracture, scoliosis, aliasing, bronchiectasis, and bifid rib
- 92.9%
- Detection rate of nodule
- 93.8%
- Detection rate of fracture
- 90.1%
- Detection rate of tuberculosis
- 94.4%
- Detection rate of calcified foci
- 89.7%
- Detection rate of others
Intelligent Diagnosis for Hepatic Space-Occupying Lesions
- Hepatic space-occupying lesions detection, with focus on textural feature analysis
- Six hepatic space-occupying lesions: primary hepatocarcinoma, hepatic metastatic tumor, hepatic cyst, hepatic hemangioma, focal nodular hyperplasia and hepatic abscess
- Four features: corona enhancement, mosaic architecture, envelope and nodule-in-nodule
- 91.3%
- Detection rate of lesions <3 cm
- 99.7%
- Detection rate of lesions >3 cm
- 18%
- False-positive rate
e) Intelligent Screening for Mammary Gland Molybdenum Target X-ray
- Enhanced detection rate of micro-calcification of molybdenum-based mammography target
- 94.8%
- 0.05mm - 1mm detection rate
- 96.3%
- Accuracy of benignity/malignancy prediction
- 14%
- False-positive rate
Intelligent Screening for Lung Cancer Pathological Image
- China's first pathological assisted screening system for lung cancer based on deep learning
- The ability to load digital pathology slides of up to 200,000*200,000 pixels, automatically detect lung-cancer cells and explore millions of data characteristics
- 91%
- Detection of lung cells
- 8%
- False-positive rate
