These features, as well as a handful of others, are being worked on over this summer.Īnother project underway for digiKam is improvements to the software's face recognition neural network engine.
The software will pin Important face tags in the sidebar and automatically assign face tag icons for a faster, more visual workflow. Users will soon be able to tell digiKam to ignore certain faces using the Reject button. Other improvements are coming to the face detection tool and related workflow, as well. In cases where the algorithm isn't trained, detected faces are instead marked as 'unknown.' Though the workflow for this updated face recognition function remains the same, the digiKam team notes that users will need to train the neural network to recognize the faces of the people they often photograph by tagging them in multiple images. The digiKam team says that testing the updated feature with 'huge collections' revealed a high level of accuracy resulting in few false positives. The neural network model now used for digiKam's face detection feature can detect faces in a variety of states and arrangements, including profile shots, burred faces, faces partially obstructed by other objects like sunglasses and more.
#DIGIKAM FACIAL RECOGNITION UPDATE#
The updated digiKam adds a number of new features and bug fixes, according to the team behind the software, including an important update to the software's face-detection capability, which now utilizes deep learning technology to better detect both human and animal faces. Open source photo management software digiKam has been updated to version 7.0.0, an update that has been a full year in the making.
0 kommentar(er)
