Dairy Facial Inference Technology

An instant and comprehensive job interview for your diary cow.

DairyFIT analyzes the structures of a cow’s face to infer dimensions of innate personality that influence her susceptibility to disease.

Use Case

For a group of candidate replacement heifers, a farmer collects facial images for each cow using any standard imaging device. Variations in facial structures are quantified using our patented image analysis technology and uploaded to our cloud, where algorithms estimate her genetic value for production and health traits. Genomic testing is recommended for the best subset of these heifers. Genomic results are later added to their profiles and combined with facial biometrics to calculate the most accurate genetic estimate possible. Farmers make final replacement heifer selections based on our genetic estimates and customizable selection indices. Individualized grouping, breeding, feeding, and veterinary care plans are then generated for the optimal precision management of each cow.

Potential

Global warming, urban encroachment, water shortages, and antibiotic resistance are just a selection of the issues that dairy farmers will face while striving to satisfy global protein needs by 2050. Altogether, they send the same urgent message: we must do more with less. After 50 years of intense selection for production traits, the modern dairy cow is a Ferrari: high performing, but also high maintenance. Under all but optimal conditions, their susceptibility to disease makes such genetics more a liability than a boon. Genomic tests for health traits are now available to farmers, but if we are to breed reliability back into our cows in time to weather this building storm, we must create smarter breeding and management systems that are as robust and dynamic as the cattle we seek to create

Business Case

Dairy cattle are governed by complex networks of biological interactions, and yet data collected on modern dairy farms remains isolated across various industry platforms. We will become the middle-man that connects these disparate sources of information. By integrating data from breeding companies, farm software, milk machines, and cow sensors, our patented algorithms will provide farmers nuanced insights into breeding and management decisions that that work with the individual natures of their animals instead of against them. Our facial biometrics are the only comprehensive, noninvasive, and robust measure of animal temperament available. We'll use that competitive advantage to convince farmers to invest in our services, but create lasting value in our company through the data we collect.

Objectives:

  1. Improve the accuracy of existing genomic tests for health and longevity traits by accounting for polygenetic temperament traits.
  2. Reduce costs of genomic selection programs for health and longevity traits by developing a prescreening algorithm for targeted testing.
  3. Develop precision management plans for a cow’s unique health risks that allow for targeted use of veterinary, feed, and sensor technology.

Team Dairy FIT

Fort Collins, CO, United States of America

What better way to celebrate being selected as finalists than a weekend-long data binge?

Fertility is the biggest risk to a cow's longevity in the producing herd, and accounts for roughly a quarter of all cullings. Here's a sneak peak into the results a weekend full of data crunching. Genetic estimates for conformation and milk producing ability on their own account for less than 5% of the variability in cow conception rate. A preliminary model that utilizes our facial metrics accounts for over 50% of the variability in the genetic estimate of cow fertility!

Think. Create. Ideate! / Published January 15, 2017 by Kaysie Jennings

Think. Create. Ideate!

It took a while for us to control the chaos of our ideate workflow. Ideas were bouncing off of every wall it seemed. We were finally able to focus everything down to the users of Dairy FIT and what we believed to be important aspects for each of them.

Free parking! Ideas welcome! / Published January 15, 2017 by Kaysie Jennings

Free parking! Ideas welcome!

Our vision of idea parking was relatively simple. We utilized whiteboard space and post-it notes to write down as many ideas as possible and categorize them somewhat based on the overall topic. These post-its could include a vague topic, a picture, a fun fact, or any thing else that the Dairy FIT crew found interesting or inspiring. The process of generating initial ideas allowed us to see how one another thought and use that information to collectively move forward as a team.

Plans for world domination... or world iteration / Published January 15, 2017 by Kaysie Jennings

Plans for world domination... or world iteration

Our plan for iteration has been thoroughly thought through and sketched out. Due to the nature of Dairy FIT, iteration should be relatively simple with cooperation from producers and industry groups. We plan to use several data sets that cover both conventional and organic farms from across the dairy to continually refine and introduce robustness into our algorithms. Further down the line, we hope to explore implementation of machine learning techniques to continuously tweak our algorithms based on data uploaded to our cloud.

Helping Dairy FIT perform to its full potential / Published January 15, 2017 by Kaysie Jennings

Helping Dairy FIT perform to its full potential

Here's one last glimpse into our design lab! Looking around the indoor Dairy FIT design lab, there are a number of features, but the most popular feature in the design lab is most likely our tea kettle, "Old Faithful." Coincidentally, "Old Faithful" is also the hardest worker in the design lab, keeping the Dairy FIT crew motivated and well-caffeinated throughout the project!

Our Team

Our Mission

We strongly believe that, when you employ the right algorithms, animal welfare doesn't have to be a zero-sum game. We aim to combine the power of new machine vision and data analytics technology to simultaneously improve the health, wellbeing and productivity of intensively farmed animals. Decades of selective breeding have allowed us to maximize the milk producing capacity of the modern dairy cow. Now, we need to ensure that these highly prolific animals are healthy enough to live up to their genetic potential. A cow's temperament plays a large role in helping her to become a successful milking animal. Temperament traits are highly heritable, but they cannot be predicted genomically, and are not effectively quantified with behavioral observations and tests. Our algorithms use facial biometrics to provide farmers an instant and comprehensive analysis of a cow's innate personality, and combine this information with additional data about her management and genetics to predict her lifetime productivity and flag potential health risks.

Our Background

Team Dairy FIT features a geneticist, an animal sensor specialist, and a behaviorist turned data junkie. Normally, our skill sets would be isolated in distinct sub-disciplines of animal science, but we are daring to cross lab-lines in an effort to create more comprehensive solutions to the ASAS grand challenge problems in animal health. Though we approach production problems very differently, we all share the same goal - ensuring that farmers and their animals have the resources necessary to continue making progress towards a hunger-free world where everyone has the opportunity to enjoy safe, nutritious, and delicious animal-based products. We see animal wellness and welfare as an opportunity, not a market threat, and believe that data-driven technologies hold the key to hacking animal temperament and realizing massive gains in the productive potential and efficiency of intensively farmed animals around the world.

Our Badges

  • Power Team
    Your team has at least three members
  • Design Lab Experts
    Attended at least one of the weekly TFF Design Lab “office hours” sessions.
  • We Famous
    Your project has been covered by your university or other media outlet
  • Get #Social
    Created team Facebook or Twitter page and attached to team page
  • Fan Club
    You have 100 likes or followers on your team Facebook or Twitter
  • Interior Architect
    Posted three pictures of your team workspace
  • Idea Parkers
    Shared a picture of your teams unique “idea parking”
  • Global Foodies
    Posted three facts about global food issues learned in the Explore mode
  • Insight + Insight + Insight
    Shared three new insights your team formulated in the Define mode
  • On Time, On Target
    Shared your Design Statement created during the Define mode
  • Post-it Fiends
    Shared pics of the chaos of the post-it filled Ideate work mode
  • Brain Food
    In the Ideate intro video, Richard shares an important tip while eating an apple. Share a team-apple-eating pic!
  • Expert Input
    Shared three pieces of expert feedback on your concept
  • GIF Masters
    Shared a GIF of the prototype building frenzy
  • Live Testing
    Shared a video of a user testing your concept
  • Way To Go
    Shared a picture of your plan for iterating your project
  • Etch-A-Sketch
    Shared your logo sketching session on your team blog
  • WE DID IT!!
    Submitted Final Pitch, including videos and concept details
The information contained here represents student project ideas developed as the result of brainstorming activities during Round 1 of the TFF Challenge. It does not represent any final business plans or commercial products.