Some treasures can only be found by experts, such as doctors or veterinary specialists. Is that so? Our customer vetevo put this to a test: They needed a fast and reliable way to detect different kinds of worm eggs in microscope images. After training and integrating our software, they are now getting a response within a matter of seconds from colabel and can provide much faster service to pet owners.
Implementing our solution further led to a perceptible quality increase as human errors when analyzing dozens of samples each day were greatly. Vetevo put colabel at the core of its most sought-after product and got rewarded with reduced cost, raised quality and improved service levels.
vetevo is building Europe’s leading pet healthcare platform, helping pet owners provide the best care for their pet’s health.
Pet owners can use vetevo to digitally manage their pet’s health (e.g., document age, weight, symptoms, vaccinations), receive personalized reminders, information and guidelines, and find, book and buy first class pet healthcare solutions straight from the vetevo app. Besides convenience, this is a more cost-effective way than the traditional, purely analog way of file documenting, searching information and going to the veterinary for all kinds of health matters.
One of their leading products is a deworming kit for home use, activate it in the vetevo app and send it back to a vetevo laboratory. The vetevo app keeps the user informed about the status of the test and asks him to input health information for more accurate test results. Once available, the lab results are displayed in the vetevo app, along with treatment and medication recommendations – a game-changer for the entire industry.
In order to recommend the right deworming treatment, two factors are relevant:
- Type of worm eggs
- Number of worm eggs
Traditionally, the samples were analyzed by veterinary specialists. After preparing the samples, they need to be analyzed under the microscope. This costs much time and is often prone to errors, because the human eye gets tired and can quickly overlook unknown patterns. As the company started selling more and more tests, vetevo needs technical support to deliver fast and high quality results from their lab.
The company was looking for new ways to raise the bar in quality and speed compared to traditional methods. Our customer therefore sought out a solution that was not only efficient but also more accurate than the traditional way, since the result was used to inform the pet owner and recommend appropriate treatment and medicine.
What our customer built
Given the importance of the product, vetevo had decided to invest in a microscope that was able to automatically take dozens of images of each laboratory sample. Backed by a rich image dataset, our customer decided to go for a process with two training stages:
- Stage 1: Analyze the image on content to find out if there are any worm eggs in the sample.
- Stage 2: Run the positive images through an object detection algorithm in order to count how many worm eggs there are.
Although not always necessary, sometimes it helps to filter out some of the noise in the data. Our neural networks show great performance out of the box on a number of tasks but in this specific case, it required a narrower target for the actual identification and counting.
In order to get the same performance in a production setting, these two stages would have to be carried over to the live process. Our customer decided to initially connect the process via Google Sheets and Zapier, and moved on to connecting our API once the process was final.
After successful integration, the solution returns a response within a matter of seconds after the microscope images are sent to our system. Mapped with different treatment recommendations, the company was able to achieve a high degree of automation.
Expressed in numbers, the model achieved accuracy of >0.99. Subsequently, only a small fraction of all incoming samples has to be manually checked by their laboratory staff while each of these manual interventions adds to the learning mechanism again. This not only led to significant cost reductions but also cost avoidance in the future as the company grows, as well as a better service level towards their customers – and not to forget, quicker help for their pets.
Our customer was able to build a key element of their service in an automated way. Through using colabel, a task thus far only carried out by experts could be digitized in a way that would otherwise require serious investments in deep learning engineers – a function that simply does not exist in their company and likely never will. Using colabel certainly did not make the whole project simple, but at least possible.