At the recently held Poultry Science Association LATAM Meeting, Phytobiotics launched their new AI-based phone application called Chicken Checker. This tool will allow poultry producers to detect possible disease outbreaks early. Farm workers will have to take pictures of the chicken faeces. The algorithm will then compare the pictures and give information on the health status of the flock. All this information will be available almost instantly both on the phone of the operator and the respective field technician as well as on a dashboard on computers.
We have talked to Phytobiotics` Dimitri Freitas and Cleverson de Souza as well as with Dr Ricardo Hummes Rauber from Vetinova, the leading experts who have worked on the validation of this brand-new and really unique technology.
Dimitri Freitas, Phytobiotics
Can you describe how long the whole process of testing Chicken Checker took?
The development process began in 2021, in an idea conception phase. The following year, the ideas became a plan. Initially, the goal was to develop a safe and efficient method that could correlate the photos of the droppings with the intestinal evaluations made at the necropsy. After having the method developed work in the last two years and confirming the security of the information and ensuring the app would be easy to use for users.
What were the main challenges when you have started the testing process?
Developing an innovation is a challenging process every day. Project Chicken Checker involves people from different fields, such as veterinarians, information technology professionals, statisticians, technical assistants, and farmers. Involving and engaging all people was the first major challenge.
What do you see as the main advantages of chicken Checker for the industry?
There are several benefits to information management in the protein industry. I would mention a few: monitoring a relevant parameter in real time, information history, results management (being able to anticipate trends and analyze potential risks) and another relevant point would be to optimize the work of technical support.
Cleverson De Souza, Phytobiotics
Can you give us some numbers behind chicken checker? Like the numbers of necropsies, farms, pictures..,
The protocol to create the algorithm, a total of 40 flocks were analysed, with 800 necropsies, 53307 pictures collected, but 33221 pictures used to make the first algorithm.
After the algorithm working, to testing and validate the CC at LAR, around 150 flocks, were used around 47000 pictures were collected by farmers.
What was the main challenge for validating the correlation between the pictures and the results of the necropsies?
The main challenge was correlating the results of necropsies and pictures of faeces. For this a huge protocol with a huge number of necropsies and pictures was performed. After that, the algorithm was trained to recognize the probability of the animal having a gut issue based on a picture.
How secure/valid are the results of Chicken Checker?
The initial analyses to create the algorithm were based on the statistical protocol and regressions analyses, these first steps allow a significative confidence in the results. The algorithm has around a 90% probability to identify gut issues through the picture. Besides that, the capacity to improve this accuracy is constantly up to date, since more and more pictures and flocks are evaluated.
Ricardo Hummes Rauber, Vetinova
What were your first impressions when you were asked to support the validation of Chicken Checker?
When I was first asked to support the validation of Chicken Checker, I found it to be a fascinating opportunity. The idea of integrating real-time data and technological advancements into poultry health management immediately struck me as a significant step forward for the industry. I saw the potential for it to streamline decision-making processes, improve health outcomes, and enhance productivity in a way that aligns with my focus on evidence-based strategies in animal health.
How do you see the future of Chicken Checker and what other application of Chicken Checker could you imagine?
The future of Chicken Checker looks promising, especially with the growing emphasis on precision agriculture and data-driven health management. As technology continues to advance, I believe Chicken Checker will evolve into an even more robust platform, potentially integrating predictive analytics and AI to forecast health risks and optimize intervention strategies. Beyond poultry, this technology could be adapted to other sectors of animal husbandry, such as swine, offering similar benefits in terms of disease prevention and productivity enhancement.
What is your advice to future users of Chicken Checker?
My advice to future users is to fully embrace the technology and make it part of their management strategy. The more data you feed into Chicken Checker, the more valuable insights you'll gain. Additionally, don’t be afraid to explore its full capabilities and collaborate with experts to tailor the system to your specific needs. The technology is only as good as its implementation, so ensure that you are maximizing its potential to drive better health outcomes and operational efficiency.
For more information about the Chicken Checker reach out to Dimitri Freitas at d.freitas@phytobiotics.com.br.
Свяжитесь с нашими экспертами или отправьте нам сообщение. Мы свяжемся с вами как можно скорее.
