Optical multiscale morphological processor using a complex-valued kernel

Area of Science:

• Image Processing
• Computer Vision
• Optical Engineering

Background:

• Morphological transformations traditionally use binary kernels and thresholding.
• Existing methods can be computationally intensive and sensitive to image imperfections.

Purpose of the Study:

• To present an alternative approach for morphological operations using complex-valued kernels.
• To enhance information processing capabilities and system robustness in image analysis.

Main Methods:

• Utilizing a complex-valued kernel with odd symmetry for morphological operations.
• Implementing a scale-space representation by continuously varying kernel size.
• Developing an optical system for morphological filtering.

Main Results:

• The complex-valued kernel processes constant image regions in parallel, increasing efficiency.
• The scale-space representation provides robustness against noise and spatial non-uniformities.
• The proposed optical system effectively performs morphological filtering.

Conclusions:

• Complex-valued kernels offer an efficient and robust alternative for morphological image processing.
• The scale-space approach enhances system resilience to image degradations.
• The developed optical system demonstrates practical application of the proposed method.