Identification Process of a Dental Implant Visible on an Input Image by Means of at Least One Convolutional Neural Network

The process according to the preceding claim, wherein classification (CLASS) comprises a comparison step of each probability of corresponding to a technical characteristic specific to an implant with a set of technical characteristics of reference implants so as to obtain for each reference implant a correspondence score.

The process according to any one of the preceding claims, comprising a learning step (APP) of at least one neural network from a database of annotated data (BDD1) and radiographic reference images, said database of annotated data (BDD1) comprising references of the reference implants and coordinates of reference regions of interest of the radiographic reference images, said learning (APP) for identifying on the reference images a region of interest and technical characteristics specific to each reference implant visible on each reference image and vectors representative of each reference image.

The process according to the preceding claim, wherein the learning step (APP) of at least one neural network comprises a step for augmenting data consisting of applying to each radiographic reference implant image at least one random transformation from the following transformations:

• rotation, trimming, enlarging, modification of brightness, modification of contrast, mirror, perspective, noise.