The Most Interesting Use Cases of 2020: Part I
Mindy Support actualizes many different projects from various industries every year. While some may be more complex than others, each project has its own nuances and individual specifications from the client that must be followed. Let’s look back at the previous year and share with you some of the most interesting data annotation use cases that we worked on.
Differentiating Good Caviar From the Bad
Caviar lovers know the exact taste black caviar should have and due to its high price, it has become associated with pure luxury. In fact, one pound of black caviar can cost around $3500. Since it is so expensive, companies need to implement a lot of quality assurance measures to make sure that only the highest quality caviar is harvested. Our client was working on a machine learning-powered system that could detect whether the caviar was harvested correctly by analyzing its color.
For example, if the caviar was even slightly gray, it could not be sent to the consumers. Our data annotators labeled thousands of images with semantic segmentation to make sure the color of the caviar was exactly right. This is one of the most detailed and time-consuming types of data annotation and we were able to complete all of the work within the specified time frames.
Training Machines to Harvest Crops
As robots are starting to be used more often in the agriculture industry, farmers are relying on them to accomplish an increasing amount of redundant and time-consuming tasks. One of them is harvesting ripe peppers. The robot needs to know the color a ripe pepper has and distinguish it from other shades of that color.
We assembled a team of data annotators to perform this job for the client. This annotation work also required semantic segmentation since there are so many possible shades of red, green, and yellow peppers. We provided our team with the needed training on how to annotate the images and how to distinguish ripe peppers from the rest. The client was very happy with the results since we allowed them to achieve significant time and cost savings.
Contouring the Roofs of Houses
Housing roofs can come in many different shapes sizes. Common forms include gable, M-shaped, hip and valley, and many others as well. Also, all of these roofs can be made out of many different materials. Our client was working on a machine learning solution that would be able to classify all of this information. For example, it would be able to look at a picture of a certain house and determine that it has an M-shaped roof made out of tiles. We assembled a team of data annotators that contoured the roofs of houses on thousands of images and labeled the materials the roofs were made out of. The client provided very detailed instructions on how the annotation work should be done and we had to hire and train our data annotators to perform the job correctly the first time. All of the annotations were done correctly and on-time due to our meticulous training methodology and rigorous QA processes.
Mindy Support Can Handle Even the Most Complex Data Annotation Projects
The use cases we described above are only a small selection of the many projects that we are actualizing on a daily basis. They are a good example of how we are able to adapt our processes to the specific needs of our clients and how we can actualize even the most innovative projects. We have 2,000 employees in six locations all over Ukraine which allows us to assemble even the most sizable teams quickly and we can also scale your team without sacrificing the quality of the annotation work. Contact us today or browse through our case studies to find out why companies from startups all the way to Fortune 500 and GAFAM companies trust us with the data annotation and BPO needs.
January 22nd, 2021 Mindy News Blog
Talk to our experts about your AI/ML projectContact us