Highlights of digiKam 8.5 include several enhancements and improvements to the Face Management feature thanks to the integration of deep-learning models into the face pipeline for faster and more accurate face detection and recognition. YuNet is now used as the default model for face detection, offering more configuration options to enhance overall accuracy and support for fine-tuning face detection for specific scenarios. On the other hand, SFace is now used as the default deep-learning model for face recognition.