Related Experiment Video
Updated: Jul 7, 2026
12:26
Control of Cell Adhesion using Hydrogel Patterning Techniques for Applications in Traction Force Microscopy
Published on: January 29, 2022
Projection-based image registration in the presence of fixed-pattern noise
S C Cain1, M M Hayat, E E Armstrong
1Dept. of Electr. and Comput. Eng., Dayton Univ., OH 45469-0245, USA.
Summary
A new projection-based method enhances image registration accuracy and efficiency, outperforming traditional 2-D cross-correlation techniques, especially in noisy conditions. This approach offers improved noise rejection for better image analysis.
Area of Science:
- Image processing
- Computer vision
- Signal processing
Background:
- Traditional 2-D cross-correlation methods for image registration struggle with fixed-pattern and temporal noise.
- Gradient-based shift estimation is computationally efficient but sensitive to noise amplification.
Purpose of the Study:
- Investigate a computationally efficient projection-based image registration method.
- Improve performance over traditional 2-D cross-correlation techniques in noisy environments.
- Enhance noise rejection capabilities for better image analysis.
Main Methods:
- Transforming images into row and column vector projections.
- Estimating horizontal and vertical shifts using 1-D cross-correlation on vector projections.
- Developing a figure-of-merit based on signal-to-noise ratio (SNR) to quantify noise rejection.
Main Results:
- The projection-based estimator significantly reduces computational complexity compared to 2-D methods.
- Improved performance in the presence of temporal and spatial noise was observed.
- The projection-based method demonstrates superior noise rejection capabilities.
Conclusions:
- The novel projection-based method offers a computationally efficient and robust alternative for image registration.
- This technique provides enhanced performance and noise resilience over traditional cross-correlation methods.
- Validated through simulations and real image sequence tests, the method is suitable for noisy image analysis.
