Computation of the normalized cross-correlation by fast Fourier transform

Area of Science:

• Computer Vision
• Image Processing
• Signal Processing

Background:

• Normalized cross-correlation (NCC) is crucial for template matching in various applications.
• Rapid computation of NCC is essential for real-time systems.
• Existing NCC computation methods have limitations in speed efficiency.

Purpose of the Study:

• To develop a novel scheme for computing normalized cross-correlation (NCC).
• To enhance the speed efficiency of NCC computation using the fast Fourier transform (FFT).
• To provide a competitive alternative to existing NCC computation techniques.

Main Methods:

• Development of a computational scheme utilizing the fast Fourier transform (FFT).
• Application of the FFT-based scheme to normalized cross-correlation (NCC) computation.
• Comparative analysis of the proposed method against existing techniques for speed.

Main Results:

• The proposed FFT-based NCC computation scheme demonstrates significant speed advantages.
• The method shows favorable speed efficiency compared to traditional techniques.
• Performance may exceed existing methods under specific search conditions.

Conclusions:

• The FFT-based approach offers a highly efficient method for NCC computation.
• This technique is particularly beneficial for time-sensitive template matching applications.
• The developed scheme presents a viable and potentially superior alternative for NCC calculations.