A sub-Riemannian model of the visual cortex with frequency and phase

Area of Science:

• Computational Neuroscience
• Computer Vision

Background:

• The primary visual cortex (V1) processes visual information through specialized neurons.
• Understanding the computational principles of V1, particularly simple cell receptive fields, is crucial for visual processing models.

Purpose of the Study:

• To develop a novel computational model of the primary visual cortex (V1).
• To introduce an image enhancement algorithm based on the proposed V1 model.

Main Methods:

• Modeling V1 simple cells using Gabor functions, incorporating orientation, frequency, and phase selectivity.
• Interpreting V1 as a fiber bundle with intrinsic orientation, frequency, and phase variables.
• Developing an image enhancement algorithm utilizing these intrinsic image properties.

Main Results:

• The Gabor function naturally induces the model's geometry and horizontal connectivity patterns.
• The image enhancement algorithm effectively utilizes orientation, frequency, and phase information.
• Experimental results demonstrate the algorithm's efficacy in image enhancement.

Conclusions:

• The proposed V1 model provides a robust framework for understanding visual processing.
• The developed algorithm offers an effective method for image enhancement by exploiting multi-feature information.