Dog Identification System Using Nose Print Biometrics
Meo Vincent C. Caya, Emmanuel D. Arturo, Chezjon Q. Bautista · 2021 IEEE 13th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM) · 2021
Advancement in technology has improved identification systems by integrating biometrics as a form of reliable authentication. Besides human biometrics, biomarkers are also present in nature and throughout various species of animals. Dogs are known to have unique patterns on their nose analogous to human fingerprints. This paper reports developing and implementing a system designed to recognize dogs with their nose prints as biometrics. The identification system employs various image processing techniques developed in past research papers. You Only Look Once (YOLO), and Scale-Invariant Feature Transform (SIFT) extract an image’s unique vital points and descriptors. These unique descriptors, along with the dog’s information, are saved for future matching. The matching is done with the Fast Library of Approximate Nearest Neighbor (FLANN). The system was able to identify 18 dogs out of 20 correctly. A total of 20 comparisons were made, making the overall accuracy of the system 90%.