Iterative Annotation to Ease Neural Network Training: Specialized Machine Learning in Pathology
Human teachable histology image annotation AI with user-friendly interface.
Automated diagnosis in medical imaging has been hindered by a lack of well-annotated images in large data sets, mainly due to the limited availability and high cost of clinical experts' time. Neural network detection of objects in these images is a potential solution to this problem, but current technologies still require numerous images annotated by human experts before they perform to a reasonable degree of accuracy. This Catch-22 situation makes the cost of creating computer vision diagnosis software prohibitively expensive.
To address this problem UB researchers designed a user-friendly annotation interface which allows intuitive, greatly accelerated neural network training. Clinicians train the software by annotating a small number of images. The software then uses this data to predict and overlay annotations on a few new images, which the user can easily correct via 'clicking and dragging' color-coded object boundaries. After only a few rounds of iterative training, a highly accurate trained AI can be produced and deployed. The software has broad applications and has already been applied to kidney histology slides and MRI images of human prostate.
- Semi-automated image annotation and segmentation
- Intuitive user interface
- Rapid AI development and deployment
- Medical Imaging
Copyright & Know-how
Technology validated in lab, TRL 4
Available for license
Nat Mach Intell., 2019 February