Revopoint researchers have been developing an image completion algorithm to be used in 3D face printing and lately they obtained exciting results.

The algorithm is used to complete the 3D face models scanned and output by Revopoint’s 3D face scanner Facense. The four-camera 3D face scanner takes photos of the face from four different angles to acquire a 3D face model from left ear to right ear at accuracy up to 0.1mm. The whole process, including scanning and processing, takes just 15 seconds or so, and users obtain a face model supporting STL, OBJ and PLY formats.

Normally, these formats are compatible with most 3D printers and ready for printing. However, since Facense can only capture the front face, the rest part of the head is incomplete. To complete the model for 3D printing, users can turn to more 3D scanners to scan the top and the back of the head, but it’s time consuming and difficult to operate and stitch the models together. To resolve this problem, Revopoint upgraded its algorithm and made the process much easier and faster.

How it works?

The AI completion algorithm matches the 3D face data (from left ear to right ear) with Revopoint’s head model database for and fuse them together to reconstruct the full head.

Revopoint offers a great number of hairstyles for users, and they can select one and instantly preview the result. After the hairstyle is picked, the 3D model can be output and printed out without further processing.

In addition, with Revopoint’s 3D bust model database, users can select the styles they prefer and apply it to the head model.

This image completion technology has lately been released as part of 3D face printing solution. It saves time and money for customers who purchased Facense, and also those who need comprehensive and personalized 3D face printing solutions. With the completed 3D head models, users can print a bust or even a whole-body statue.