Gradient-based 3D skin roughness rendering from an in-vivo skin image for dynamic haptic palpation

Area of Science:

• Dermatology
• Computer Graphics
• Medical Imaging

Background:

• Palpation is crucial in dermatology but poses risks like infection or damage.
• Indirect palpation via haptic rendering of 3D skin models offers a safer alternative.

Purpose of the Study:

• To develop an accurate 3D skin surface reconstruction method for haptic rendering.
• To enable virtual palpation of skin conditions using single in-vivo images.

Main Methods:

• A novel method using simple gradients and bilateral filtering for 3D skin surface reconstruction.
• Preservation of skin roughness, wrinkles, and textures in the generated 3D polygonal meshes.

Main Results:

• The proposed method accurately reconstructs 3D skin surfaces from single images.
• Evaluations using ground-truth surfaces and clinical images (acne, herpes simplex) demonstrated superior accuracy compared to existing methods.
• Objective metrics like Hausdorff distance confirmed the method's effectiveness.

Conclusions:

• Accurate 3D skin surface reconstruction is essential for effective haptic palpation.
• The developed method provides precise 3D skin surface reconstruction from single images, preserving surface roughness.