The SecondOpinion® system consists of three parts: in-office application or UserInterface ("UI") (the "SecondOpinion® Client"), Application Programing Interface ("API"),and Computer Vision.

The SecondOpinion® Client is continually monitoring a local (or networked) resource where dental radiographs (refer to Section 2.4 for list of supported image types) are stored. Once new imagery is discovered, the Second Opinion® Client invokes cloud-based APIs which submit the imagery to Computer Vision Models ("CV Model", “CV Models”) for processing. The metadata that these CV Models produce describe the nature and location of detected pathologic or non-pathologic features that may appear in the radiographs.  

This metadata information is then sent back to the Second Opinion® Client, which renders it for visual display and review within the Client’s user interface. Any potential pathologic or non-pathologic features detected will appear inscribed within color-coded boundary boxes overlaid on the original radiograph.

The entire process outlined above takes five to ten seconds to complete.

SecondOpinion® CADe detects potential pathologic or non-pathologic features based on their visual appearance. Features are detected because they closely resemble known features present in the radiographs used to train the Second Opinion® CV Models.

The system can assist the radiologist in minimizing observational oversights by identifying regions that may warrant a more meticulous review.

The Second Opinion® software may be used only as an aid for detecting certain potential pathologic or non-pathologic features that can appear in bitewing and periapical radiographs.It is not for use as an aid in radiographic diagnosis or interpretation.