Eyeglasses lens contour extraction from facial images using an efficient shape description

Area of Science:

• Computer Vision
• Image Processing
• Biometrics

Background:

• Accurate extraction of eyeglass parameters from facial images is crucial for applications like virtual try-on and facial recognition.
• Existing methods often struggle with variations in eyeglass styles, lighting conditions, and facial poses.

Purpose of the Study:

• To develop an automated system for precise extraction of eyeglass position, frame shape, and lens size from facial images.
• To introduce novel methods for eyeglass shape representation and contour extraction.

Main Methods:

• Utilized Fourier descriptors for an original model of eyeglass lens shape representation.
• Developed a Fourier morphing technique to generate a comprehensive search space from limited representative lens shapes.
• Implemented a multi-stage Monte Carlo sampling algorithm for accurate lens contour extraction.

Main Results:

• The proposed system demonstrates effective and accurate extraction of eyeglass position and lens contours.
• Experimental results validate the robustness and precision of the developed algorithms.

Conclusions:

• The novel approach using Fourier descriptors and Monte Carlo sampling significantly advances automated eyeglass parameter extraction.
• The system offers a reliable solution for applications requiring precise eyeglass information from facial images.