Cross-Modal Image Matching Using Calibration Mediators

The patent introduces a calibration image containing known feature points as an intermediary reference. This calibration image serves as a mediator between color and infrared images, enabling the system to establish accurate correspondence relationships despite differences in sensor types and field of view. The calibration process creates a mapping model that compensates for the inherent mismatches between different image modalities.

The patent transforms the matching problem by changing the parameter space. Instead of directly matching features between color and infrared images with different characteristics, the system converts both images into a common reference framework using calibration data. This parameter transformation allows accurate matching by aligning feature points through calibrated transformation matrices that account for sensor-specific variations.

The patent performs calibration as a preliminary action before actual authentication. By pre-establishing the relationship between sensors with different fields of view using a calibration image, the system creates a reference model that enables accurate alignment during authentication. This preliminary calibration step ensures that subsequent matching operations can achieve high precision despite the inherent field of view differences.

The patent creates a virtual copy of the calibration relationship that can be applied to multiple sensor pairs. The calibration model, once established, serves as a reusable template that captures the geometric and optical characteristics of the sensor arrangement. This copied calibration data enables consistent alignment across different authentication sessions without requiring repeated physical calibration.