
digiKam 8.7 adds AI face scan improvements, new GPU options and AI auto-rotation plugin
digiKam 8.7.0 arrives as the newest release of the open source photo management and editing software. AI-powered face detection now automatically initiates a new recognition scan whenever users confirm or tag new faces, making tagging more efficient. When users reject an AI-suggested face match, digiKam records the rejection and immediately presents the next likely match if available, streamlining the face-tagging process.
Additionally, a new option enables users to check for GPU availability within the artificial intelligence processing pipelines. When supported, images are sent to the graphics processor for resizing and color conversion, and then processed by the neural network through OpenCL. Users can selectively disable OpenCL for reference engine models, while a dedicated test feature in the configuration dialog verifies GPU compatibility for AI tasks.
Alongside these changes, a newly introduced plugin uses deep learning to perform automatic orientation detection, allowing users to apply the correct image rotation and fix large batches in one step. Internal components have also been refreshed: the Libraw decoder and ExifTool library have both been updated, and the G’MIC-Qt plugin in the editor and batch queue manager is now at version 3.5.0. While Linux and Windows bundles upgrade to Qt 6.8.3, macOS gains Qt 6.9.0. This release also addresses several notable bug fixes.