Rotation and illumination invariant interleaved intensity order-based local descriptor

Area of Science:

• Computer Vision
• Image Processing
• Pattern Recognition

Background:

• Local intensity ordering patterns enhance image matching discriminability.
• High dimensionality of traditional descriptors limits practical image matching due to time constraints.

Purpose of the Study:

• To propose a novel descriptor that reduces dimensionality and matching time.
• To maintain or improve matching performance compared to existing methods.

Main Methods:

• Introduced the interleaved order-based local descriptor (IOLD).
• Constructed descriptors by considering neighboring sample points in an interleaved manner.
• Explored local intensity order patterns in a multi-support-region scenario.

Main Results:

• The IOLD significantly reduces descriptor dimension and matching time.
• Achieved comparable or superior matching performance against prominent descriptors.
• Demonstrated robustness to complex illumination differences and noise.

Conclusions:

• The IOLD offers an efficient and effective solution for image matching.
• The interleaved approach addresses the dimensionality and time complexity issues of traditional descriptors.