A Multi-Task Deep Learning Framework for Pose-Invariant Face Recognition and Facial Landmark Localization

Key points

• Face recognition systems often experience performance degradation under large pose variations due to changes in facial appearance and visibility of discriminative features.1
• To enable simultaneous optimization of identity recognition and facial structure learning, the original Cross-Pose Labeled Faces in the Wild (CPLFW) dataset was adapted into a multi-task learning dataset by associating facial images with identity labels and five-point landmark annotations.1
• The proposed framework consists of two task-specific branches operating on a common facial input: a ResNet50-based branch for identity recognition and an EfficientNetV2B0-based branch with spatial attention for facial landmark localization.1
• The results indicate that combining identity classification and landmark regression within a single optimization framework provides an effective approach for learning robust facial representations under cross-pose conditions1

Sources (1)

• [1]A Multi-Task Deep Learning Framework for Pose-Invariant Face Recognition and Facial Landmark Localization
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 02:23 PM

  This study presents a multi-task deep learning framework that jointly addresses pose-invariant face recognition and facial landmark localization within a unified architecture.

  Face recognition systems often experience performance degradation under large pose variations due to changes in facial appearance and visibility of discriminative features.

Extractive summary: sentences quoted from the sources.