Our Technology

Facial Inference Technology (FIT)

There is a depth of antiquated livestock knowledge suggesting that aspects of an animals temperament that impact their training and management can be inferred by reading their face. A moose nose will be friendly. A wide forehead will be reliable. The FIT algorithmic pipeline pairs the power of modern machine vision with purpose-built machine learning tools to bring these old cowboy adages into the 21st century. The result is an instant job interview for your cow that can be tailored to your unique production environment at a fraction of the cost of traditional genetic evaluations or genomic testing.

Patents:

Image Analysis for Determining the Characteristics of Animals

Image Analysis for Determining the Characteristics of Individuals

Image Analysis for Determining the Characteristics of Pairs of Individuals

Image Analysis for Determining the Characteristics of Groups of Individuals

Livestock Informatics Toolkit (LIT)

The LIT toolkit is a suite of interpretable machine learning algorithms built specifically for applications with precision livestock technologies. Optimized for knowledge discovery over prediction, fundamental principals of animal biology and behavior have been infused throughout this analytical pipeline, allowing the LIT toolkit to recover patterns from these data streams at the level of the smallest farms or even large pens. This provides farmers a means to engage with their data and extract insights that are uniquely suited to your unique farm environment.

Publications:

Mind the Queue: A Case Study in Visualizing Heterogeneous Behavioral Patterns in Livestock Sensor Data Using Unsupervised Machine Learning Techniques

Livestock Informatics Toolkit: A Case Study in Visually Characterizing Complex Behavioral Patterns across Multiple Sensor Platforms, Using Novel Unsupervised Machine Learning and Information Theoretic Approaches

Biometric Informatics Toolkit (BIT)

The BIT toolkit is a suite of machine vision algorithms purpose-built for applications in quantifying livestock conformation and morphology. Heavily inspired by traditional approaches to visual inspection of livestock quality, the BIT toolkit uses projective geometry to break down complex morphological features into interpretable and reliable 1D morphometrics with statistical properties that are optimally-suited for use in downstream statistical analyses.

Publications:

Improving the Reliability of Scale-Free Image Morphometrics in Applications with Minimally Restrained Livestock Using Projective Geometry and Unsupervised Machine Learning