Robust Point Set Matching for Partial Face Recognition.

Area of Science:

• Computer Vision
• Biometrics
• Pattern Recognition

Background:

• Traditional face recognition relies on complete facial images.
• Real-world scenarios often involve occluded faces, hindering recognition.
• Existing methods struggle with partial face identification.

Purpose of the Study:

• To develop a robust partial face recognition method.
• To enable accurate person identification from incomplete face images.
• To address challenges posed by occlusions in unconstrained environments.

Main Methods:

• Keypoint detection and local textural feature extraction from partial faces.
• A robust point set matching algorithm integrating textural and geometrical information.
• Calculating face similarity based on aligned feature set distances.

Main Results:

• The proposed method effectively recognizes individuals using partial face images.
• Experimental validation on four public datasets confirms the approach's effectiveness.
• The method demonstrates superior performance compared to traditional holistic approaches in occlusion scenarios.

Conclusions:

• The developed partial face recognition technique is effective for unconstrained environments.
• Integrating local textural and geometrical features enhances recognition accuracy.
• This approach offers a viable solution for face identification when full facial data is unavailable.