What have we achieved with AI in Clinical Imaging?
MRI
Classification of MRI Brain images into abnormal and normal, as good as Radiologists. We are working on individual disease classification to create a complete diagnostic system. This is expected to help Radiologists manage heavy workload.
Volume calculation of hippocampus and cardiac ventricles from MRI images.
Classification of normal and abnormal inter-vertebral discs, as good as Radiologists. Working on detailed description of discal abnormalities.
Classification of normal and torn anterior cruciate ligament (knee), as good as Radiologists
CT
Detection of intracranial hemorrhage on CT Brain, as good as Radiologists.
Classification of lung nodules into benign and malignant, better than Radiologists.
X-rays
Classification of health check-up x-rays as normal and abnormal, as good as Radiologists.
Detection of pulmonary tuberculosis and pneumothorax on chest x-rays, as good as Radiologists.
Detection of bone fracture, better than Radiologists.
Bone age calculation from hand x-ray, better than Radiologists.
Ultrasound (US)
Classification of thyroid ultrasound into normal and abnormal, and further classification of diseases on abnormal scans, as good as Radiologists.
Classification of ocular (eye) ultrasound into normal and abnormal, and further classification of diseases on abnormal scans, as good as Radiologists.
Please check this page often for updates.
We are working on creating modality/body area specific complete diagnostic systems. This is expected to help Radiologists manage heavy workload and reduce error rates especially in the night.
