The digiKam team has announced the release of digiKam 8.5, a maintenance update for the popular open-source photo management software. This version enhances several features, particularly focusing on Face Management, which has been upgraded to incorporate deep-learning models. This integration improves the speed and accuracy of face detection and recognition.
With the new update, YuNet is adopted as the default model for face detection, allowing for enhanced configuration options that can be tailored to specific needs. SFace is now the default model for face recognition. Notably, the updated face-matching algorithm requires user confirmation of just one face before it can proceed to recognize other detected faces.
In addition to face management improvements, digiKam 8.5 also enhances color label usability. Users can now utilize a linear gradient color bar beneath image thumbnails to identify important items quickly.
Other noteworthy features in this release include enhanced support for large TIFF file formats, refined date and time formatting, a new option to prevent pixel blurring in high zoom during preview, support for new Apple extensions with the HEIF file format, usability improvements to the progress manager, and better open-source calculations support through OpenCV.
Under the hood, the release incorporates an updated Libraw RAW decoder, an upgraded ExifTool for metadata management, and an improved Exiv2 shared library.
You can find more details on the release announcement page and download digiKam 8.5 from the official website as a portable AppImage. For those using Flatpak, it is available on Flathub.