Introduction
Preview Detection AI deep learning solution in 3DHISTECH Scanner Software enables scanners to operate autonomously by recognizing the sample area, thereby speeding up scanning, improving focusing quality, and saving storage space.
The training samples were collected from the widest possible range, however, there may be a sample type that is underrepresented in the 3DHISTECH dataset and is not recognized by the algorithm in perfect quality. You can report such preview images to the development team using the form below.
Feedback Process
1. Collect preview images
The Preview image can be saved as a simple JPG image from the SlideViewer/ClinicalViewer software. The more samples are fed back, the more accurate training can be achieved for new types of samples. Only the Preview JPG images are required, the entire slide does not need to be uploaded, so these images will not contain patient data or identifiers. To achieve improvement usually a few tens or hundreds of images are sufficient.
2. Upload preview images
Collected preview images can be uploaded on the following page:
