Muzzle-Based Cattle Identification Using Artificial Intelligence (AI) For Tracking & Tracing In Bangladesh

Abstract

A big barrier to insurance companies offering cow insurance is the absence of cattle identification systems. This lack of technology had devastating financial consequences for marginal farmers as they did not have the opportunity to claim compensation for any unexpected events such as the accidental death of cattle. Using machine learning and deep learning algorithms, adorsho praniSheba has solved the bottleneck of cattle identification by developing and introducing a muzzle-based cattle identification system. It is scientifically proven that the muzzle of a cattle is unique, resembling a human fingerprint. This is the fundamental premise that prompted us to develop a face recognition like system to extract the uniqueness of a cattle muzzle. For this purpose, we collected about 20,000 images from 1,000 cattle. Contrast-limited adaptive histogram equalization (CLAHE) with sharpening filters were applied in the preprocessing step to remove noise from images. We used Facebook's devised YOLO algorithm for cattle muzzle detection in the image, and Google devised the FaceNet algorithm to directly learn embeddings from muzzle images. Our system works with an accuracy of 95.41%, at a remarkably low false acceptance rate of 0.003 with an F1 score of 92%. Thus, adorsho praniSheba has elegantly solved the problem of cattle identification and ensured livestock farmers easy access to cattle insurance. As a consequence, formal lenders and investors are now assured of their investments as our technology is now underwriting their finance. This has ushered in a new era for our rural economy.

Full text

Title: Muzzle-Based Cattle Identification Using Artificial Intelligence (AI) For Tracking & Tracing In
Bangladesh
Safayet Khan, Data Scientist, Machine Learning Projects, adorsho praniSheba Ltd
Email: safayet@pranisheba.com.bd, Cell # +880 1937-711055
Abstract(250 words):
A big barrier to insurance companies offering cow insurance is the absence of cattle identification
systems. This lack of technology had devastating financial consequences for marginal farmers as
they did not have the opportunity to claim compensation for any unexpected events such as the
accidental death of cattle. Using machine learning and deep learning algorithms, adorsho
praniSheba has solved the bottleneck of cattle identification by developing and introducing a
muzzle-based cattle identification system.
It is scientifically proven that the muzzle of a cattle is unique, resembling a human fingerprint. This
is the fundamental premise that prompted us to develop a face recognition like system to extract the
uniqueness of a cattle muzzle. For this purpose, we collected about 20,000 images from 1,000 cattle.
Contrast-limited adaptive histogram equalization (CLAHE) with sharpening filters were applied in
the preprocessing step to remove noise from images. We used Facebook's devised YOLO algorithm
for cattle muzzle detection in the image, and Google devised the FaceNet algorithm to directly learn
embeddings from muzzle images. Our system works with an accuracy of 95.41%, at a remarkably
low false acceptance rate of 0.003 with an F1 score of 92%.
Thus, adorsho praniSheba has elegantly solved the problem of cattle identification and ensured
livestock farmers easy access to cattle insurance. As a consequence, formal lenders and investors
are now assured of their investments as our technology is now underwriting their finance. This has
ushered in a new era for our rural economy.
Keywords/Topics:
Artificial Intelligence, Machine Learning, Deep Learning, Image Processing, Computer Vision, Big
Data, Cattle Identification, Muzzle Detection, Agrotech, Fintech, Insurtech
Co-Authors:
Dr Rezaul Alam Reza1, Md. Mahadi Hasan Sany2, Fahim Hossain Sifat3, Fida Haq4, Imtiaz
Rahi5and Dr Md Aolad Hossain6
1, 2, 3.4 adorsho praniSheba Ltd,Dhaka, Bangladesh; 5, 6 shurjoMukhi Ltd, Dhaka, Bangladesh