System and method for mobile fingerprint recognition
AssigneeThales Dis France SAS
InventorsChinmay Sahu, Songtao Lester Li, Amit Pandey
Life of this patent
Estimating the term…
Abstract
A system (100) and method (600) for mobile fingerprint recognition is disclosed. The method (600) comprises receiving an image of a fingerprint. The method (600) further comprises detecting a set of parameters associated with the fingerprint. Further, the method (600) comprises aligning the fingerprint in the image based on the detected set of parameters. The method (600) further comprises extracting a normalized fixed size Region of Interest (ROI) from the aligned fingerprint. Further, the method (600) comprises generating a fixed-length feature set from the extracted ROI. The method (600) further comprises performing one or more operations on the fingerprint based on the generated fixed-length feature set.
Claims13
A system for mobile fingerprint recognition, the system comprising:
- a processing engine coupled to a memory, wherein the processing engine is configured to execute instructions stored in the memory to:
- receive an image of a fingerprint;
- detect a set of parameters associated with the fingerprint;
- align the fingerprint in the image based on the detected set of parameters;
- extract a Region of Interest (ROI) from the aligned fingerprint;
- generate a fixed-length feature set from the extracted ROI;
- and
- perform one or more operations on the fingerprint based on the generated fixed-length feature set.
A system for mobile fingerprint recognition, the system comprising:
- a processing engine coupled to a memory, wherein the processing engine is configured to execute instructions stored in the memory to:
- receive an image of a fingerprint;
- detect a set of parameters associated with the fingerprint;
- align the fingerprint in the image based on the detected set of parameters;
- extract a Region of Interest (ROI) from the aligned fingerprint;
- generate a fixed-length feature set from the extracted ROI;
- and
- perform one or more operations on the fingerprint based on the generated fixed-length feature set;
- wherein to perform the one or more operations, the processing engine is configured to:
- generate a multi-indexed feature representation using a local fingerprint database and a global fingerprint database;
- perform a similarity comparison of the extracted fixed-length feature set using a deep-learning-based matching algorithm;
- and
- perform at least one of score-level fusion or hybrid matching.
A method for mobile fingerprint recognition, the method comprising:
- receiving, by a processing engine, an image of a fingerprint;
- detecting, by the processing engine, a set of parameters associated with the fingerprint;
- aligning, by the processing engine, the fingerprint in the image based on the detected set of parameters;
- extracting, by the processing engine, a Region of Interest (ROI) from the aligned fingerprint;
- generating, by the processing engine, a fixed-length feature set from the extracted ROI;
- performing, by the processing engine, one or more operations on the fingerprint based on the generated fixed-length feature set;
- wherein performing the one or more operations comprises:
- generating, by the processing engine, a multi-indexed feature representation using a local fingerprint database and a global fingerprint database;
- performing, by the processing engine, a similarity comparison of the extracted fixed-length feature set using a deep-learning-based matching algorithm;
- and
- performing, by the processing engine, at least one of score-level fusion or hybrid matching.
A non-transitory computer-readable medium storing computer-executable instructions for mobile fingerprint recognition, the computer-executable instructions configured for:
- receiving, by a processing engine, an image of a fingerprint;
- detecting, by the processing engine, a set of parameters associated with the fingerprint;
- aligning, by the processing engine, the fingerprint in the image based on the detected set of parameters;
- extracting, by the processing engine, a Region of Interest (ROI) from the aligned fingerprint;
- generating, by the processing engine, a fixed-length feature set from the extracted ROI;
- performing, by the processing engine, one or more operations on the fingerprint based on the generated fixed-length feature set;
- wherein performing the one or more operations comprises:
- generating, by the processing engine, a multi-indexed feature representation using a local fingerprint database and a global fingerprint database;
- performing, by the processing engine, a similarity comparison of the extracted fixed-length feature set using a deep-learning-based matching algorithm;
- and
- performing, by the processing engine, at least one of score-level fusion or hybrid matching.
Dependencies are inferred from the claim text.