Hierarchical Fingerprint Image Merging System

The patent segments the merging process into multiple hierarchical levels. Instead of directly merging all images into one, it creates intermediate merging results at different levels (first level merging, second level merging, etc.), where each level processes a subset of images. This segmentation allows error isolation and prevents error propagation across the entire set of fingerprint images.

The patent implements a nested hierarchical structure where merging operations are organized in nested levels. The first level merges individual fingerprint images into intermediate results, the second level merges those intermediate results into higher-level results, and so on. This nesting allows the system to build up reliable merged images incrementally, with each level benefiting from the error-correcting properties of the previous levels.

The patent combines multiple smaller fingerprint images captured by a compact sensor into a single comprehensive merged fingerprint image. By using hierarchical merging algorithms that align and integrate features from multiple scans, the system achieves the feature completeness that would otherwise require a larger sensor area, thus avoiding increased device complexity and cost.

The patent transitions from a single-dimension approach (one large sensor capturing all features at once) to a multi-dimensional approach (multiple smaller sensors or repeated scans capturing different regions, then merged). This dimensional change allows the system to achieve complete fingerprint coverage through temporal and spatial multiplication rather than through a single large sensor area.

The patent performs preliminary alignment and feature extraction on individual fingerprint images before the merging process. By pre-processing each image to identify key features and establish coordinate transformations, the system reduces the computational complexity of the actual merging operation, enabling faster processing of multiple scanned images and reducing overall enroll time.

The patent creates intermediate copied representations of fingerprint images at different hierarchical levels during the merging process. These copied intermediate results allow the system to efficiently combine information from multiple scans without repeatedly processing the original large datasets, thus reducing computational time while maintaining complete feature capture.